Hacker News Reader

You are misrepresenting the conclusions of that report. The report clearing distinguishes between "mitigation" and "suppression" in the second paragraph.

> However, the resulting mitigated epidemic would still likely result in hundreds of thousands of deaths and health systems (most notably intensive care units) being overwhelmed many times over. For countries able to achieve it, this leaves suppression as the preferred policy option.

The aggressive social distancing implemented in the U.S. falls under "suppression," and therefore is not considered in the predictions you quoted:

> suppression will minimally require a combination of social distancing of the entire population

Two paragraphs after the bit you quoted:

> We therefore conclude that epidemic suppression is the only viable strategy at the current time. The social and economic effects of the measures which are needed to achieve this policy goal will be profound. Many countries have adopted such measures already, but even those countries at an earlier stage of their epidemic (such as the UK) will need to do so imminently.

The report predicted 5,600-48,000 deaths in Great Britain if all four interventions discussed were put in place. The UK has had ~26,000 deaths.

This prediction for GB (suppression) is 80-98% lower than the prediction you quoted (mitigation). A naive application of these ratios to the US suggests that the predicted number of deaths with all interventions would be something like 24,000 - 220,200. The U.S. is currently at ~60,000 deaths.

So, erm, no - the United States has not "beat every single expert pre-peak total deaths prediction; by a lot."

No, GB's excess mortality compared to last year is already greater than 35 thousand death, and that's not taking into account the 4-day lag between the actual death and the report (unlike "covid19" hospital death. GB is not testing home sudden death (that are often caused by cardiac arythmia that might be induced even in very mild covid19 cases), so only excess mortality data is reliable:

https://www.euromomo.eu/graphs-and-maps/#z-scores-by-country

[edit] That said, your argument still stand and in fact is even better with this data

Uk exess mortality is already at 35 000, quite close to the "successfull suppression" plan that put the death toll at 40 000.

"First, note that ~60k is how many people died and tested positive AND were coded as COVID-19. It's not necessairly how many people died because of a COVID-19 infection. We will not know the real #'s until all-cause death is examined. There appears to be a significant drop in "normal" pneumonia deaths for example. It also ignores the error bars on the testing false/positive/negative rates."

Do you have a website like euromomo for excess mortality in the US? Covid19 causes arrhythmia in 7% of cases, even very mild ones [0] and arrhythmia is the main cause of sudden death for people over 40 years old (i have ectopic heartbeat so i'm a bit concerned with this)

[0] https://www.uptodate.com/contents/coronavirus-disease-2019-c...

> If the authors had given a table for the US, we could see what their model predicted, but they did not, and _you_ assuming that the US is a linear relation to GB may or may not be accurate.

The Imperial College team actually did model country-by-country numbers in a follow up report on March 24th:

https://www.imperial.ac.uk/mrc-global-infectious-disease-ana...

The United States estimates for several different scenarios are:

Unmitigated: 2,654,410 deaths

Mitigation strategy (social distancing whole population): 1,358,164 deaths

Mitigation strategy (enhanced social distancing of elderly): 1,090,267 deaths

Suppression strategy (triggered at 1.6 deaths per 100,000): 474,227 deaths

Suppression strategy (triggered at 0.2 deaths per 100,000): 84,124 deaths

The differences between mitigation and suppression are explained in the paper.

Thank you. Any estimate on when we crossed the supression strategy death/100k threshold?

The triggers considered were 0.2 deaths / 100k population / week and 1.6 deaths / 100k population / week. For the U.S. that corresponds to ~700 deaths / week and ~5,200 deaths / week.

The United States as a whole hit the first trigger (according to reported deaths from https://github.com/nytimes/covid-19-data/blob/master/us.csv) sometime around 3/19, and second trigger sometime around 3/27.

replying to your reply via edit:

> _you_ assuming that the US is a linear relation to GB may or may not be accurate.

Right, this is why I called it 'naive.'

My main point stands: your comment misuses the prediction that does not assume full suppression as if it applied to the current situation of aggressive social distancing.

> The US beat every single explicitly enumerated via a integer expert pre-peak total deaths prediction; by a lot.

From 'nodamage's comment pointing to the ICL's "Report 12," they predicted 84,000 deaths for the U.S. in a situation where suppression measures were put in place around 3/19.

Given that the U.S. reported number of deaths is at ~61,000 and still seeing 1,500+ deaths per day, this doesn't exactly make me want to claim victory against the estimate - more like 'thanks for the predictions, ICL team, for helping convince people and governments to shut down.'

The ~60k could be an overestimate, yeah. Maybe it's counting people that have coronavirus and are then hit by a truck, or people who would have died of pneumonia regardless of covid19.

But it could be the other way around, too: https://www.nytimes.com/interactive/2020/04/28/us/coronaviru....

https://www.cdc.gov/nchs/nvss/vsrr/covid19/index.htm

> COVID-19 death counts shown here may differ from other published sources, as data currently are lagged by an average of 1–2 weeks.

> *Data during this period are incomplete because of the lag in time between when the death occurred and when the death certificate is completed, submitted to NCHS and processed for reporting purposes. This delay can range from 1 week to 8 weeks or more, depending on the jurisdiction, age, and cause of death.

Indeed. As this tails off, it's going to be the official spot to check the prediction[1] linked here against.

https://news.ycombinator.com/item?id=23028036

Also: https://mobile.twitter.com/johncardillo/status/1256577527662...

Neat, so come back in a few months.

Also, I haven't got a clue who John Cardillo is, but that tweet doesn't really say much other than he thinks something is a scam. The 37k currently reported at the CDC link corresponds to the deaths reported for ~12 days ago at https://github.com/nytimes/covid-19-data/blob/master/us.csv, so it isn't clear that the discrepancy is not largely explained by the lag to process death certificates.

EDIT:

I looked at some of his other tweets, and it seems like he is upset that the count here from the National Vital Statistics System https://www.cdc.gov/nchs/nvss/vsrr/covid19/index.htm doesn't match the count here https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/case..., and interprets that to mean that the CDC is lying to the U.S. population. I disagree with that assessment, and the CDC explains the NVSS data here: https://www.cdc.gov/nchs/nvss/vsrr/covid19/tech_notes.htm:

> Provisional death counts in this report will not match counts in other sources, such as media reports or numbers from county health departments.

John Cardillo says that "You’ve been misled while lunatics try to destroy your county and steal your way of life" (presumably he means "country"?), the CDC explains the discrepancy. I'll go with the CDC.

I agree. Per capita we are doing exceptionally better than most European countries. My European friends think it’s like a warzone over here but outside of pockets in NYC and smaller pockets around the country the hospitals aren’t full. Whether or not that is due to our response or certain characteristics of our nation (e.g. very spread out, lots of space) is up for debate, sure. But we are doing very well.

Exceptionally? The US is better than some European countries, but significantly worse than others https://ourworldindata.org/grapher/total-covid-deaths-per-mi...

Also if it is true that there had been significant community spread early on, thenthe US number is strongly underestimated. If not the US is on a different part of the curve and behind the deaths in Spain or Italy.

>> The US beat every single expert pre-peak total deaths prediction; by a lot.

This is mostly because the UW IHME model was brutally terrible as was the Imperial College modeling. They predicted insane amounts of death that had no chance of happening while baking in social distancing orders.

We massively overreacted and the convenient out is "well, we beat the models, that's a good thing!"

Not if the dose was more poison than antidote. I could have predicted 5 million deaths pulling it out of my ass and I shouldn't get credit for the US coming in under that estimate.

"We massively overreacted"

I agree, but that was not an accident. Cui bono?

Seems worth repeating: "We can't let the cure be worse than the problem itself"

https://twitter.com/GPIngersoll/status/1255564315891032064

I don't think it was malice in this case, though I get your point. I think the people that did the models were just incompetent and also shaded them to be overconservative because they felt it would be better (so slightly fraudulent, perhaps).

But all evidence points to the IHME to just be terrible at their jobs and influenced a country to massively over-react.