How risky is your social meeting? A new calculator.

September 3, 2020 8:24 AM

We reviewed published research about COVID, and used it to make rough estimates about the risk level of various activities in microCOVIDs. I've been doing this math on my own to check my local rates, but this is way easier and better. Included: Location impacts, mask type impacts, indoors/outdoors impacts, and your personal risk rate impact. Shows work and based on science.
posted by bbqturtle (30 comments total)

It turns out, my weekly dungeons and dragons group in Minneapolis is a 1% chance each week... but that goes down to 0.05% if outdoors. At the very least, we are going to start opening windows / sliding doors!

I found this useful to see the impact of various changes (Would it be better if we wore masks, or if we went outside?)
posted by bbqturtle at 8:27 AM on September 3, 2020 [1 favorite]

https://twitter.com/PatJWu/status/1299918900381937665

A twitter user uses this template to evaluate going to the gym. In my opinion, his conclusion that an additional 1% chance a year of getting covid by attending weekly is too high, is within acceptable tolerances (especially since going to the gym helps with immune response)
posted by bbqturtle at 8:34 AM on September 3, 2020

This is helpful to contextualize risk in my area...I wish it were even more granular by location, but it still makes my head less swimmy.
posted by wellred at 8:43 AM on September 3, 2020 [2 favorites]

Ughhh god damn it if I don't get to go to the gym for another year I am going to start SETTING FIRES.
posted by We put our faith in Blast Hardcheese at 9:06 AM on September 3, 2020 [2 favorites]

Ughhh god damn it if I don't get to go to the gym for another year I am going to start SETTING FIRES.

If you wear a mask, do it alone, and set the fire outdoors, burning the place the fuck down represents a less than 1% chance of exposure to covid.
posted by maxwelton at 9:11 AM on September 3, 2020 [40 favorites]

They seem to be making a lot of assumptions in what they are taking from the research, but I don't have the background to actually contradict any of the work. At the very least it seems like they are making very conservative choices, which is good, as the main worry for a project like this would be overconfidence in its estimates.
posted by Think_Long at 9:14 AM on September 3, 2020

It turns out, my weekly dungeons and dragons group in Minneapolis is a 1% chance each week...

And if D&D has taught us anything, it's that a 1% die roll for something awful to happen will occur about three times in ten.
posted by FatherDagon at 9:18 AM on September 3, 2020 [44 favorites]

Some of my family members insist on having people over for dinner on weekends—inside, and unmasked. Not a huge number of people, 10 or less, but we've politely declined their invitations since this whole thing began.

We insist on seeing those family members outdoors, and with masks on.

It's nice to see that our concerns are well-founded, because often we're made to feel as if they are not.
posted by vitout at 9:25 AM on September 3, 2020 [20 favorites]

Interesting. So walking through this, my much-missed weekly beer-share dinners would be low-to-moderate risk depending on how excited people get (I'd be alright with 20% of a week's risk budget on it) since everybody's grocery-shopping & otherwise podded up; but there's really no getting around the bus ride blowing that out of the water.

😞 About what I expected, but that really underscores both how restricted things have to be to make "a 3-hour unmasked indoor dinner" work and how far we have yet to go where I'm at.
posted by CrystalDave at 9:44 AM on September 3, 2020 [4 favorites]

The case data was not remotely accurate for the two counties I tried.
posted by HotToddy at 10:47 AM on September 3, 2020 [3 favorites]

I like having more tools to understand the risks of our behavior and encourage us to behave more responsibly, or at least more in line with our individual risk tolerances*. Before taking the results of the calculator as gospel, there's a lot to read here, and by the authors' own disclaimer:

We have read a lot of experts' research, but we are not ourselves experts in this topic.

I'd be really interested to see comments on this analysis by some people who are experts, whether in infectious disease, epidemiology or statistics. Or, to compare this to kind of model to comparable work by experts. As a relatively informed—and hopefully relatively intelligent—layperson, I think I can recognize things that are clearly absurd or contradicted by substantial data, but I don't have much reason to believe I can spot good-faith mistakes or unintentional bias in work produced by smart laypeople. Certainly experts can differ, can have their own agendas, or can be manipulated by influential people, but it seems wise to always consider the opinions of people with knowledge and experience in the discipline(s) in question. Just because someone is a genius and an expert in one field, doesn't automatically make their work in other realms similarly credible.

* "Base jumping for some, taking a brisk walk for others" — The Simpsons, lightly paraphrased
posted by Cogito at 10:57 AM on September 3, 2020 [1 favorite]

Ughhh god damn it if I don't get to go to the gym for another year I am going to start SETTING FIRES.

(a) We're full on fires here, thanks.
(b) If you set fires, then you can't do ANYTHING outside.
posted by jenfullmoon at 11:08 AM on September 3, 2020 [9 favorites]

I like the idea of this tool as a way to think through risk - something I am having trouble doing without collapsing in a pile of stress.

The case numbers seem correct for my state and county but the % positive rate is wildly inaccurate (higher than its ever been).
posted by entropyiswinning at 11:28 AM on September 3, 2020 [1 favorite]

I'm sure if the % positive rate or case numbers are incorrect, the authors would be very interested. You can post on their github with comments of that nature. Is it possible that the numbers are a week behind, or from a different source than you usually use?
posted by bbqturtle at 11:31 AM on September 3, 2020

(a) We're full on fires here, thanks.

Well where I am there are no fires yet, and for a fire to travel from Chicago to California would be...impressive enough that I'd almost like to see it? I mean I wouldn't half-ass it, I'd Mrs. O'Leary's cow some shit, but even so...

Honestly I'm just at the point where if it's less than 100% chance of acquiring covid I'm prolly just gonna do it. Nobody else is being careful anymore and I'm fucking sick of wrecking all my mental and physical health so some dipshit can do shotskis in Lincoln Park.
posted by We put our faith in Blast Hardcheese at 2:00 PM on September 3, 2020 [1 favorite]

Interesting - an unmasked outdoor socially distanced (more than 6 feet) meetup in my Southern state with 2 friends is a 40 in 1 MILLION chance of catching COVID. Those seem like pretty good odds to me, but it's considered moderate risk? Am I just being innumerate here?
posted by Brain Sturgeon at 2:05 PM on September 3, 2020 [1 favorite]

@BRain Sturgeon: Its relative to the risk tolerance dropdown right at the bottom of the page. If you choose 1% per year, it allocates ~0.02% as allowable each week. Which happens to be about 200 in a million. So your one-time meetup used 20% of your "risk allowance" for that week. And it looks like anything between 10% and 50% of your weekly allowance is marked as moderate, presumably because it doesn't take many of them to exceed your budget.
posted by Jobst at 2:22 PM on September 3, 2020 [7 favorites]

Okay, thanks, Jobst, that was helpful. I kind of like the idea of a coronavirus budget. Unfortunately, being married to front line worker means I'm constantly broke :(.
posted by Brain Sturgeon at 2:30 PM on September 3, 2020 [5 favorites]

Okay, thanks, Jobst, that was helpful. I kind of like the idea of a coronavirus budget. Unfortunately, being married to front line worker means I'm constantly broke :(.

That was my takeaway, too... I'm supposed to go into the office once a week, so I used up my budget for two weeks right there today. And my partner works in a clinic, so that's.... 30 weeks budget there?

Idk, I feel like 1% is a pretty unrealistic risk tolerance for most folks in the US at this point, unfortunately. If I can be careful enough to get my chances down to 5-10% of getting out of 2020 without getting this thing I think I'd call that a success under the circumstances.
posted by geegollygosh at 5:31 PM on September 3, 2020 [4 favorites]

I also think the idea of a risk budget is an interesting concept. Though that's probably not how we live our lives and manage risk generally, maybe we should?

I think two concepts which are highly useful for perspective are micromorts (basically a 1-in-a-million chance of death, but easier to do math with than 0.000001) and knowing your baseline probability of dying within the next year based on your age and gender.

Take a 40-year-old woman in the US. She has an 0.001429 probability of dying within the next year. That's about 14 in 10,000 or 1,429 micromorts. To decide how much risk one is willing to take on, we can convert the risk of contracting COVID-19 into the risk of dying from it to give perspective. If one targeted the 1% annual risk of contracting the virus mentioned above, and assumed a 1% infection fatality rate*, that 's 0.0001 probability of dying from COVID-19 or 100 micomorts: a net increase of about 7%. That's the same increase in risk one would get from driving 23,000 miles (a little under 2 years worth for the average driver in the US), or a bit less than the risk of death from giving birth.

Now, what are we willing to give up in order to keep our risk that low is a very individual question, but I think it's important to not give more weight to risks merely because they're novel.

* Probably an overestimate, but we're also ignoring the potentially negative but nonfatal effects of contracting the virus.
posted by Cogito at 9:06 PM on September 3, 2020 [3 favorites]

I foolishly plugged in scenarios that my Florida-dwelling family keep putting themselves into, and now I'm really regretting it. I thought for half a second of sending this to them so they could play with the numbers and see for themselves, but I worry that a "1% risk" would translate to "perfectly safe" to them.

I feel like this entire COVID season has been one prolonged exercise in trying not to freak out, again and again, over the stupid risks taken by people I love.
posted by DingoMutt at 9:54 PM on September 3, 2020 [3 favorites]

I think it would be interesting to swizzle this tool into a collective action calculator: if we all acted like this, would the R_e be below 1.0? Or what if they acted like that?

Or else if we bake in people's various work exposure as immutable (it shouldn't be, but God Money may not be fixable in 2020), and make it: if we all in our 'discretionary' actions acted like this what would that do?
posted by away for regrooving at 11:50 PM on September 3, 2020 [3 favorites]

Now, what are we willing to give up in order to keep our risk that low is a very individual question

What's unusual about COVID risk is that it is not an individual question. When my risk of infection increases, my risk of infecting others increases as well. The micromort calculations you made do not factor in the micromorts inflicted on the (R=?) people I would infect, and also the micromorts delivered to everyone in the downstream chain of transmission. We can probabilistically calculate the burden they would bear, but there's a question of consent, for the people downstream did not consent to my participating in a high risk activity.

Say what you will about base jumpers, they're not risking the lives of unwilling strangers.
posted by justsomebodythatyouusedtoknow at 12:47 AM on September 4, 2020 [9 favorites]

The moment you and everyone else spends from their risk budgets there will be inflationary pressure on the banked mico-covids because the rate of infection increases potentially exponentially. Given America's lack of substantive outbreak control (ie. monetary policy) America will be the Wiemar Republic of Covid. You're banked micro-covids will end up being close to worthless.

I'd also like to point out that in addition to Micro-Covids and Micro-Morts you probably need to include Micro-Morbs (morbidity as in long term disease/disability) given the mounting evidence of heart problems.
posted by srboisvert at 11:58 AM on September 4, 2020 [6 favorites]

but we're also ignoring the potentially negative but nonfatal effects of contracting the virus.

But why are we, though? I'm far, far less worried about dying from covid than I am about being horribly sick for a month or more, having a terribly un-fun hospital stay, having long-lasting or potentially permanent lung damage, and/or passing it on to someone else at higher risk of complications (or really anyone). These are all far more likely outcomes than outright dying for someone my age and all of these will be counted in the "recovered" statistics...but they're all more than bad enough to be worth a few precautions for me.

The original site focusing on the risk of catching covid rather than the risk of dying from it is a feature, not a bug.
posted by randomnity at 1:17 PM on September 4, 2020 [7 favorites]

So I tried calculating the risk my friend who lives in AZ takes when she goes to work each day at the grocery store, and it came back as "Dangerously high." I don't think I'm going to tell her that, but it's pretty devastating.
posted by WalkerWestridge at 2:48 PM on September 4, 2020 [5 favorites]

What's unusual about COVID risk is that it is not an individual question. When my risk of infection increases, my risk of infecting others increases as well.

That's an excellent point. I'd say that not spreading to others is the most important thing for us to do individually to achieve the greatest benefit collectively. And I think the best strategy for not spreading to others is not getting oneself infected. If one notices symptoms or becomes aware of a high risk experience, one should get tested, and self-quarantine until there's evidence against an extant infection. I don't think any of that is controversial, is it?

So, I think the question of how best to not infect others really does come back to individual choices aimed at avoiding one's own infection. I think if everyone successfully achieved a 1% annual infection probability*, we'd drive cases to zero quite effectively. In the US there have been ~6 million confirmed cases since January 13th. So, in about 64% of a year, 6 million of the 328 million people have had confirmed cases or roughly 3% on an annual basis (assuming trends continue linearly). Of course, that's confirmed cases; the actual number is much higher, but I think it's clear that if we individually achieve a low probability of infection that collective action achieves the benefit for everyone.

* Given the wide gulf between confirmed cases and actual (potentially asymptomatic) infections I'm not sure how to interpret the idea of 1% risk of contracting COVID-19. Presumably that's 1% of actually contracting it, as opposed to 1% chance of having a confirmed case.
posted by Cogito at 10:04 PM on September 4, 2020

but we're also ignoring the potentially negative but nonfatal effects of contracting the virus.

But why are we, though?

For simplicity's sake. We don't yet have enough data to even have a solid IFR—which is why I used the upper bound estimate of 1%—and we have even less data about associated morbidity. That's not to say that's not hugely important stuff, but to do a proper analysis, one would need to be able to estimate the impact of something that incorporates both morbidity and mortality like quality-adjusted years of life.

What I was trying to get at in my previous comment was a perspective for thinking about risk in terms of activities we've dealt with our whole lives, and simplifying things in terms of micromorts made that workable in a way that I hope is still valuable. I definitely don't mean to discount the very real nonfatal outcomes of a COVID-19 infection and as a member of a cohort that has a relatively low probability of death, I worry about those kinds of things quite a lot.
posted by Cogito at 10:14 PM on September 4, 2020

Tomas Pueyo (Thread reader)
After countries applied a Hammer, many didn't learn to Dance. With the same conditions emerge the same results. Until they learn to Dance, they will face ongoing cases, deaths, lifelong conditions, and a hurt economy.
posted by adamvasco at 6:28 AM on September 5, 2020

I have long understood that the Hammer is a dance, one so splendid that it cannot be touched.
posted by justsomebodythatyouusedtoknow at 11:50 AM on September 5, 2020