Thread by @DellAnnaLuca

THREAD: VIRUS RISK FACTORS

Scenario: yesterday, I got a coffee with one of my best friends.

Here are some of the factors that influenced our risk of catching COVID.

1/N

2/ We decided to go to a small coffee shop, with 6 people inside.

We could have gone to Starbucks (60 clients), but by going to a coffee with 1/10th of the people, we cut our risk in 10, all other things equal.

2B/ (Actually, a bit less, because some of the clients in the Starbucks would have been too far away to infect us with the same likelihood of those around us. So, let's downgrade the risk reduction from 1/10 to 1/5.)

3/ Yesterday, COVID prevalence in Italy was 1.6k per million people; it was 18 per million last June.

This means that, all other things equal, this one coffee was as risky as a hundred coffees last summer.

4/ We stayed a hour. Because clients come and go, we saw twice as many clients than if we stayed half an hour, therefore roughly doubling our risk.

5/ We kept our mask on for the whole chat, except the minute in which we drank our espresso. This means that we didn't breath unprotected the air of the (say) twenty clients that came in afterwards. That roughly halved our risk.

6/ If everyone kept their mask on except the minute they drank the coffee, assuming an average stay of 15 minutes per client, we would have cut our risk by (I estimate) a fifth.

As per all the other numbers in the thread, it's a subjective estimate based on real phenomena…

7/ In this case, on the one hand, people would have exhaled without a mask only 1/15th of the air; on the other hand, masks don't fully capture exhaled virus nor the relationship "viral load / risk" is linear. Hence my "1/5" guesstimate. For sure the ratio is lower than 1.

8/ The green pass law allow only the vaccinated inside coffee shops.

The vaccinated are contagious for less time. Based on studies I've seen, the integral of viral load seems to be 25% lower in the vaccinated; thus I estimate that this measure reduced our risk by roughly 25%.

9/ Note that studies differ a lot in the estimation of the above. As with the rest of the thread, what matters is not actual numbers, more the reasoning.

10/ Assuming that a few groups of friends contained a mix of vaccinated and unvaccinated, and the green pass rule constrained them in staying outdoors, I expect that it led to fewer people staying inside, further reducing our risk.

11/ Both my friend and I work from home, don't attend large social events, very seldom meet indoors, and wear N95 most of the time while indoors.

That massively lowered the risk that the person I was taking the coffee with was positive. I estimate 90% (again, guesstimate).

12/ Two days before, I spent the new year's evening in a tiny gathering (me, my wife, and a friend) and my friend did similarly (5 people). If we had gone to a large social event, the risk for us and everyone else in the coffee shop would have been proportionally higher.

13/ More in general, my risk and the risk I pose for others is dependent on my and others' interactions, not just between us but with our own individual networks.

A meeting of "work-from-home" people is less risky than a meeting of "in-person" workers, all other things equal.

14/ The choice of going to the small café whose clients are mostly locals also reduced our risk compared to the Starbucks near the train stations, whose customers are more likely to have met multiple people from multiple networks.

15/ The coffee shop had mediocre ventilation and no HEPA filter. Had there been better ventilation, the risk for everyone would have been lower.

16/ This was an incomplete review of risk factors regarding meeting a friend over a coffee. It's a thought experiment whose usefulness comes not from the numbers involved (they're guesstimates) but the reasonings.

17/ In particular, risk factors are multiplicative. This means that two different coffee shop meetings might have wildly different risks

A 15-minutes June coffee in a shop with six clients might be 2000 times! less risky than a 60-minutes January meeting in a shop with sixty ppl

18/ A breakdown of the different risks during a coffee with a friend:
- 60 vs 15 minutes → 4x risk
- June vs January → 100x risk
- 60 vs 6 customers in shop → 5x risk

4 x 100 x 5 = 2000

(explanation for the coefficients in the thread above)

Wildly different risks!