[ISL]

Cornell is testing ~everyone once a week. After an initial small flurry of cases, they've picked off 1-2 cases a day before they have a chance to enter the broader populace.

https://covid.cornell.edu/testing/dashboard/

Their case-positivity rate is ~100x lower than many places in the United States.

[ollien]

I go to a university where we're doing something similar - weekly testing for everyone at minimum, 2x per week if you're high risk or live in a dorm with a communal bathroom. For the most part, we've been doing well, but recently we've had a small surge (~100 students in quarantine, about 4-10 cases per day, with about 4000 undergrads, only part of which are on campus at the moment).

We're in a much more suburban area than Ithaca, so that probably plays a factor.

[bluedino]

The problem with testing once a week is there are four more days you can be positive on.

Also, by the time you get your test results, you have been positive and spreading.

I get tested on Monday mornings, got a call Thursday night telling me not to come in on Friday. What good did that do?

[ISL]

In most of the US, you'd just keep going to work until you felt sick, because you never got tested. If you were asymptomatic, you'd keep right on going to work.

One need not catch every case in order to make the epidemic burn out. One needs to reduce the number of transmissions/case below one. As we begin to win, contact-tracing becomes increasingly effective. When multiple-contact tracers can focus on a single known case, it is possible to really dig in and track cases effectively.

[jlokier]

It does a lot of good. Every little intervention really helps, even if it's seems far from perfect.

It's because the system as a whole is balanced on a critical threshold. It's like water boils at 101°C and condenses at 99°C.

The point is to reduce R below 1. When R > 1 the disease gets almost everywhere eventually. When R < 1 the disease fades away and most people don't get infected.

Currently R is above 1, but it's not a lot above 1. (We're lucky. There are other diseases where it's much higher.)

So even if you have been spreading on those 4 days, you are not spreading from days 5 onwards. That might be 10 days less spreading.

You are not giving it to large numbers of people in those 4 days either. As a spreader, you're adding risk to others, but only 1 or 2 will get it directly from you probably.

When you isolate on day 5, that's a reduction in the amount of spreading and it reduces R. When your traces also isolate from your day 5, that's a further reduction in the amount of spreading and R. That's recursive, if their traces also isolate, it's a further reduction, etc.

Altogether, if everyone follows these changes, there's a reasonable prospect of R falling below 1 enough for the disease to fade out. Because it's close enough to 1 already, that behaviour can make a significant difference.

In practice people do not comply, and they lie sometimes (https://www.bbc.co.uk/news/world-australia-55011790). So R does not fall as effectively as it would if there was high compliance. But every bit of compliance helps reduce R.

When R is consistently < 1 due to social changes, regular testing and behaviour changes in response (even with days of delay), tracing, social distancing and masks and workplace changes and commute changes, etc. then we will find the disease fades away.

It looks unlikely this will happen until we have widespread vaccination in some countries (USA is one of the leading examples), but the more we reduce R meanwhile, the fewer people will die or get long covid in the end. And fewer here means in the millions.

[cryptoz]

Agreed, but what good did it do? It prevented you from further spreading the virus more than if you had not got the call.

Every bit helps. Perfection is the enemy of good. Rather test weekly than never. Rather test daily than weekly. Why not hourly or every 15 minutes though? Have to have a balance somewhere.

What good does it do to get a test and have results 2 hours later, if you spread the virus in that 2 hours? Better than spreading it for 2 weeks! But it still allows for it to spread. Something is way better than nothing.