Hacker News Reader

Comment by fabian2k | Hacker News Reader

Given the current numbers in the US, it does not seem likely that you can reopen universities in any kind of safe way. Anything that puts a lot of people into small rooms is problematic, and I really can't see anything that is close to the regular operation being possible in the fall.

The only way to be able to reopen inherently risky activities like this is to suppress the virus sufficiently that you can handle the remainder with mask mandates, contact tracing and prohibiting mass events. I'm in Germany where the number of cases is drastically lower than in the US right now, and I doubt we'll be able to reopen universities fully in the fall. With the dramatic numbers from the US right now, it does seem extremely unlikely that they could be controlled enough in the fall to make any kind of reopening safe.

View on HN

My alma mater recently sent out an email of their plans to reopen campus this fall. Their reasoning is a study done by the operations research department found that even if they do classes online in fall, many students will return to campus anyways, only without monitoring and testing in place. Reopening campus would at least give the university some control over student behavior.

That's like saying that all the WFH firms should herd the employees back into the office, because their employees are out and about visiting friends and shopping for groceries, anyways.

This comparison isn't quite accurate, because we're talking about students relocating physically, which isn't as likely to happen in an employee WFH situation.

Cornell?

Given how many people have gone through the disease (nearly everywhere it is < 5%, so essentially no difference vs February) you can't really open anything anywhere safely, not without a plan for fast shutdown and accurate measurements to trigger that shutdown. Unless there's a vaccine. This is just basic SIR modelling, epidemiology 101.

Compartmental models are not typically covered in an introductory epidemiology course.

Usually not covered until an advanced methods course.

But do you disagree with the main point?

It's sound epidemiology, but I still reluctantly disagree on practical/political grounds:

1) There appears to be no serious effort to extend the initial CARE payments into the summer and fall. As such, we can't possibly close force more businesses to close and then just stand by as people lose their homes and starve. We're already experiencing significant social instability. We have to be very careful about this, not to mention to core moral issues.

2) Our country is so divided (and most of us have played a role in that, myself included) that it's impossible for the government to order such strict interventions without invoking massive backlash from a significant portion of the population. And again, back to my point about social instability. COVID-19 is far from the worst thing that could happen to the country at this point.

Can't argue with that. Damned if you do, damned if you don't.

Safety is not a binary condition. Since eradicating the virus is no longer a feasible option in most places, the risk level going forward is inevitably going to be higher than it was a year ago. Even the development of a vaccine won't fundamentally change that reality. So we're going to have to lower our expectations of safety.

> So we're going to have to lower our expectations of safety.

We really don't. In-person classes at a university are not something one should risk infection for. Yes, remote classes are not ideal. Dying, killing others, and having a lifetime of potential complications, are all worse.

> Since eradicating the virus is no longer a feasible option in most places

You have cause and effect backwards. Eradicating the virus is harder because of continued failures to maintain measures keeping it under control.

or, just let what will happen, happen? i hear a lot about case numbers, almost nothing about mortality rates

Mortality is not the only thing that matters. We don’t know what this virus does to survivors, even potentially asymptomatic ones. The little that we do know appears to be alarming (potentially permanent lung damage, some potential cases of brain damage).

We already know about viruses that sit inside people asymptomatically for years and years before wreaking havoc. Of course we cannot live permanently paralyzed by “what ifs,” but planning that focuses entirely on mortality rate is ignoring a huge, huge piece of this equation.

Many survivors are left with permanent health effects.

You may not die, but you may not be able to practice your favorite sports ever again because of permanent lung scarring.

There is no evidence that there are significant long term health implications for the vast, vast majority of people infected by this coronavirus. If you have citations that state otherwise, please share.

This shouldn’t be dismissed out of hand. Absence of evidence is not evidence of absence.

The parent comment may be referring to the growing number of “long haulers” who have “recovered” yet seem to be suffering from chronic symptoms. [1]

This may be anecdotal, but it’s a growing number of people and worth taking seriously. This is a novel virus, after all.

1: https://www.theatlantic.com/health/archive/2020/06/covid-19-...

Absence of evidence is not evidence of absence.

Not necessarily "absence", but it does provide an upper bound. There are millions of people who have survived the virus. If it was common to be incapacitated for months, we'd have more than anecdotes.

I have been ill for 4 months now.

There are stories coming out of people who have not died but certainly have not "recovered" in the sense most people would expect. See this for instance: https://twitter.com/DaniOliver/status/1279155358666305541 I don't really know how you expect students to complete their education if they're ill like that.

Also, even if that weren't the case and everyone did recover completely after two weeks, missing two weeks out of a 15-week college class is quite a bit. So even if your plan is to have your students get covid at some point in the term, have some fraction be asymptomatic, some fraction recover from it, and a couple of them die, you still need some sort of coherent educational plan for the large fraction that are going to be in the campus ICU for a couple of weeks.

I think you missed the whole reason to “flatten the curve”. The hospitalization rate is the problem, not mortality rate. 20-30% of people who get the virus end up in the hospital since there are no known treatments available. For now, if we have hospital capacity you’re very likely to survive COVID-19. If we run out of supplies and overwhelm the limited medical staff the mortality rate is going to skyrocket and we will run out of places to store the dead bodies.

20-30% of people who get the virus end up in the hospital

That's way too high. According to https://www1.nyc.gov/site/doh/covid/covid-19-data.page, 0.65% of New York City residents have been hospitalized for COVID-19. Antibody tests show at least 20% of NYC residents have been infected, so the hospitalization rate is more like 3%.

The goals are different, though. US universities don't hope to find a reopening plan that won't spread the virus at all; they just want to find a way to keep it from spreading too much.

I don’t get the impression that anyone expects anything to be 100% safe. This is a straw man. Everyone agrees the goal is to prevent it from spreading “too much,” but any proposed school reopenings (from kindergarten up to university) look like they’d land way above the line of acceptable spread.

Maybe a line we can all agree on is: If the expected resulting R_0 is above 1, it’s probably not a good plan as it will not be sustainable. Since you probably can’t know whether it’ll result in an R_0 above 1 before doing it, then in order to reopen you must have robust surveillance in place to quickly detect the R_0 and have a plan to shut back down in the case it reaches 1.

There's quibbles about whether the boundary needs to be exactly 1, but yeah, I agree with your core point. A reopening plan needs a strategy for how to scale back if something goes wrong, and it's hard to imagine how you could scale back transmission in a dorm full of college students.

Yeah, I think there’s maybe a more general solution of the form: Places can publicly declare their R_0 threshold for shutdown and their surveillance mechanism to detect when it’s hit.

I would imagine that, given a detailed surveillance mechanism that’s strictly adhered to, you could calculate the maximum “time to detection” of how soon after hitting that threshold you’d know about it.

Post those 4 things: R_0 threshold, surveillance mechanism, time-to-detection, and shutdown plan then people can choose if they want to sign up for that risk.

Of course in the education case specifically, the real crux of the issue is the differential between the educators’ (generally older) risk tolerance and those of the students/the needs of students families (in the pre-university case).

> Post those 4 things: R_0 threshold, surveillance mechanism, time-to-detection, and shutdown plan then people can choose if they want to sign up for that risk.

That's exactly what's driven me batty about governments' responses.

After the initial lockdown, that's precisely what they should have done.

Instead, we got a schizophrenic "Everything is fine" and then "Everything is danger! Surprise second lockdown!"

No wonder people are pissed. It'd be better and more effective policy to just lay out the methods, criteria, and current status.

(PS: Technically, you're talking about R_t, or current, rather than R_0, or initial)

Oof, indeed! Thank you for the correction.

Lots of things broken with public response unfortunately.

Sure! Ethan's have to watch out for each other. Just noticed your first name on this reply.

Btw (because I also peeked at other links), if another UX discovery data point re: pharma research would help, flip me an email (in profile). My father just retired from T10 drug discovery R&D (pathology), and would probably be willing to talk your ear off re: process issues.

Amazing! Thank you. Will be reaching out :)

I don't think that is possible in situations that are suitable for SARS-CoV2 transmissions. The most extreme example is probably the meat processing plants, it spread like crazy in those conditions. Another example is the choir practices that also infected a lot of the participants.

Anything that happens under the conditions suitable for superspreading events probably can't rely on any half-measures.

What specific numbers? Everything I've seen shows way more people have been exposed than though, a vast majority of them don't even generate anitbodies, and the overall fatality rates all across the board are dropping.

Just go to the beaches in Indiana and Florida and you'll see thousands of people out on the beaches, and it's been that way for a month and there's no spike in Fatalities (and please don't go on about cases. Those numbers include PCR, antibody and everyone who is getting test now for regular procedures at hospitals who are testing positive but have had no or minimal symptoms).

Reposted from a previous comment, as answering the same thought:

The danger of SARS-CoV-2, given the relatively low mortality (and based on genetic modeling of potential mutations), was never that it was going to crest with Ebola-like proportions (you get it, you die), but that it was going to crest with a magnitude that would overwhelm available medical resources.

Unfortunately, that crest is a binary proposition: either you have spare medical capacity (and serious Covid-19 patients receive high-quality care) or you have none (and serious Covid-19 patients receive no treatment).

So the danger is not death rates now. The danger is (effective reproductive number) -> (expected case count) -> (serious, hospitalized cases of Covid-19) -> (exhausted medical capacity) -> (Italy-esque spike in deaths).

Reply on news.ycombinator.com