The White House is discussing easing social-distancing guidelines as early as next week amid a broader debate over how much economic loss the country can bear to save an unknowable number of lives threatened by the novel coronavirus pandemic.
President Trump has told people that he wants to reopen the economy as soon as possible but his interest runs counter to the advice of public health experts, including in the administration, who have warned that the guidelines remain necessary.
Here are two imaginary conversations, one with a traditional health adviser named Suit, and one with an analytical health adviser named Geek.
—–1—–
President: I want to lift the social-distancing guidelines by the end of next week.
Suit: You can’t do that! The virus will spread! People will die!
President: But we’re killing the economy. I’m going to do it anyway.
Suit: I resign!
—–2—–
President: I want to lift the social-distancing guidelines by the end of next week.
Geek: You can’t do that! We need to first test a random sample in each of several geographical areas to see the distribution of the virus in the population. And we also need to run the experiment, probably a few times to get robust results. And we need to agree on benchmarks for the results that would have to be met in order to lift the guidelines.
President: What’s the point?
Geek: The random sample will tell us the prevalence of asymptomatic infected people in various geographic areas. The experiment will tell us how dangerous those people are, meaning how likely it is that they will infect others while asymptomatic.
If the experiment finds that the doorknob infection rate is less than 5 percent and the in-person infection rate for asymptomatic virus carriers is less than 5 percent, then we can safely lift many of the restrictions. Continue to encourage disinfecting and handwashing, and avoid large gatherings, but open up businesses.
If the experiment finds a result of more than 20 percent for either the doorknob infection rate or the in-person infection rate, then we should not lift the social-distancing guidelines. We would need to first meet benchmarks for treatment capacity. That might require training and equipping more health care workers, or better yet finding a reliable pharmaceutical treatment.
Suppose the experiment finds a result between 5 and 20 percent for either the doorknob infection rate or the in-person infection rate. Then decide differently for each region. If the prevalence of asymptomatic infection in a region is either very low (less than 5 percent) or very high (over 30 percent), you can lift the guidelines in those regions.
If the prevalence is currently low, then lifting the guidelines will cause the virus to spread in that region, but starting from a low baseline. So in several weeks you may have to recommend restoring restrictions in that region.
If the prevalence is high, then the people with the virus will already have encountered most of the people that they might infect (social interactions are not random). The spread rate going forward is likely to be low, as long as people stay stick to normal routines and avoid interactions with strangers or unfamiliar places.
———
Note: The numbers in the conversation are purely illustrative. I have not thought carefully about what the actual numbers ought to be.
You’re thinking along the right lines, but not with the right parameters. The recommendation should be: (1) Immediately, exempt from the lockdowns anyone wearing a mask; masks, combined with handwashing, reduce the risk of infectious spread to virtually zero to or from that person. (1) ASAP, get masks to everyone, surgical masks will do, N95 would be better.
Good ideas.
As a bureaucrat myself, that second conversation never makes it past first line supervisor, much less ten or fifteen levels. And if it magically does, if you can’t fit it in under two sentences or a quad chart, the decision makers won’t even entertain it.
What you are missing is both arguments here are the same, suit is just putting geek in decision maker interest framework.
It’s not that agency managers are stupid, they just don’t care about this stuff. You aren’t promoted in the government attempting to do the right thing or trying to understand data / make fact based decisions, if fact you are generally better off if you ignore facts period. The incentives simple don’t exist here and the geek will never get heard.
If there were such a thing as real statesmen, you might get those conversations. But we don’t have statesmen, we have modern politicians, who have different interests and ways of looking at the world.
So, in reality, Tucker Carlson drives to Mar-a-lago and tells the President facing re-election that if he doesn’t do these things, he will be the biggest loser in history. Shutting down the economy will be rough, but then again, every CEO in the country who ever said anything bad about him will literally have to come begging for his help, which will provide opportunities to exercise leverage, and also, feel amazingly satisfying.
America: 4++ weeks until peak, waiting for effective Strategy
I’m looking for four signals:
1. Grand Strategy – Flatten the Curve (score: YES)
2. Strategy – Monitor, Isolate, Trace (score: NO)
3. Tactics to support strategy (score: NO)
4. Tactic Adoption: (score: NO)
When all four signals are in place the 4++ will switch to 4?? since their is a lag in feedback time. When it looks like the peak is reached then I’ll switch to 4– to indicate four more weeks of decline until the sustained community spread nears zero.
What we are seeing now is rhetoric about the Grand Strategy without any focus on the effective Strategy required. It is true that the curve for South Korea and China took about 8-10 weeks but this is the shape of the last 8 weeks after effective tactics are widely adopted. Dr. Deborah Birx, response coordinator of the White House Coronavirus Task Force, said that sustained community transmission has been going on in New York City for about four weeks. It is a mistake to think of this curve as fixed. Without widespread effective tactics the growth will continue until about half the population is infected or recovered.
Hospital capacity collapses at a tiny fraction of any given population. Dr Birx also that that they are focused on mortality. They should be focused on respiratory related hospital admissions. It is a mistake to downplay morbidity. This disease leaves survivors with permanent asbestos-like lung damage.
Seattle is testing a representative sample
https://publichealthinsider.com/2020/03/23/introducing-scan-the-greater-seattle-coronavirus-assessment-network/
So far, estimates of US infections are in the hundreds of thousands. In other words, less than 1 in 1000 Americans is thought to be currently infected. So testing a representative sample to get +/- 5% confidence in the number of infected might take half a million tests or more. And by the time you were finished, the data would be obsolete.
We’d learn a tremendous amount even if we didn’t get to a 5% confidence interval in the number of infected. Running several point-in-time tests would give even more info, as would testing for antibodies. I’d include every person who tested positive for antibodies in each subsequent test to get stats on the decay of immunity over time.