New York Times columnist Thomas Friedman has returned to make specious analogies to China again. This time he want to institute a “more democratic version” of the contract tracing China has done:
In sum, if we are going to save the most lives while getting the most people back to work to prevent an epidemic of unemployment, depression and despair, it is going to require a federally coordinated, democratic version of the China strategy.
Let’s take that at face value. The “China strategy” was to shut down Wuhan’s public transportation, suspend train service to the city, and close the roads going into it. Short of declaring martial law, that is beyond the president’s power and the president can’t order the mayor of New York or the governor of New York State to do it.
But I have a deeper question. How do you go about tracing a person you happened to be sitting in the same car with on the subway? Let alone doing so democratically? I don’t believe there’s any way to do it other than coercively.
However, I’m skeptical about taking anything the Chinese authorities have told us at face value. More to the point what did the South Koreans do? There is a description in this New Yorker article.
Leaving aside that privacy considerations would render the approach the South Koreans took untenable, graph theory would suggest that scale is important. You’d expect the complexity to be on the order of the square of the number of people involved. In other words what is just barely practical in South Korea would be completely unworkable in the United States even if we discarded everything we’ve considered sacred in privacy rights.
I guess my point here is that rather than looking for a Chinese solution or a South Korean solution, the preferred strategy in the U. S. should be one that makes some sense here rather than urging us to remake our society so it’s more Chinese or South Korean.
“I guess my point here is that rather than looking for a Chinese solution or a South Korean solution, the preferred strategy in the U. S. should be one that makes some sense here rather than urging us to remake our society so it’s more Chinese or South Korean”
Uh, er yeah. The strategy for NYC metro, Miami -Dade, or Cook County probably differs from that for DeKalb County.
Who’d a thunk it…………..
I agree that a contact tracing strategy is not really viable given the highly contagious nature of this novel coronavirus. It may work in contained, somewhat isolated communities where ‘everyone knows your name.’ However, most of the US is in large, higher density cities where a person will come in contact with more strangers than acquaintances. Your subway analogy is spot on.
Even if there was some form of geo-tracking to confirm contacts between infected and previously uninflected individuals, what is the follow-up? More testing, perhaps multiple testings, to determine actual spread. Seems impractical that we will ever achieve the ability to administer tests on the scale that would be needed.
IMO there is a persistent view that we’re going to test everybody repeatedly without any real recognition of the cost not just in money but in time, energy, and attention that would entail.
And Manhattan Beachcomber? Cute.
Friedman phrases things the wrong way.
It is good strategy to learn and understand what works in other countries, know what does not work, and apply lessons in a culturally appropriate way.
On a long time scale, the Japanese and Chinese have done this since modernization started 150 years ago. Look up the “Iwakura Mission”.
The question is what lessons are there to learn. That has to be done without ideological blinders.
Take South Korea, here is an excellent article on what South Korea did. https://www.theatlantic.com/ideas/archive/2020/05/whats-south-koreas-secret/611215/
Here is the key phrase in the article — “Central isolation is not yet a part of most U.S. states’ response to the virus. But it ought to be. One of the hallmarks of the countries that have contained the COVID-19 threat has been the systematic separation of sick people and their contacts from healthy people.”.
Was testing the key, no. Testing is only 1 of 3/4 things that work together.
By the way, contact tracing + centralized quarantine scales up a lot.
Social networks form clusters and the virus does too. A policy of identifying clusters and isolating clusters from other clusters can get a lot of cases quickly without an extraordinary number of tests.
I agree with that. Just as I think that soliciting the contributions of experts not just in medicine and epidemiology but in engineering, public policy, etc. is a good strategy.
There’s a question that needs to be answered first. Can it be?
Thought experiment: imagine that 80% of the people you need to isolate centrally are black.
I believe that if testing can lead to being removed from your family, there will be plenty of unused tests to go around.
I have never advocated being making centralized isolation compulsory in the US. This is where culturally appropriate comes in.
Voluntary with incentives — get paid $100 a day until doctors discharge you. Included free meals and (diverse) on-site medical attention, free internet. Can bring kids along too (because kids are at low risk).
May be enough to get people interested, especially if the alternative is staying at home and likely infecting and risking their loved ones.
Half of our local cases are traced back to an employer, Smithfield foods. Pork plant. It’s about 35 miles west but many workers live here. All the workers get tested, 2 weeks paid leave if positive and asked to isolate at home. I hope they did but yesterday was Cinco de Mayo. It really looks like all employees at the plant should soon be virus positive. But there’s still the broader contacts. I doubt two weeks is enough.
Looks like the plan is to let the virus burn itself out. Dismantling the task force is a good way to turn down the spotlight on the death count and it looks like they will do that then emphasize the positives and tune out the negatives. The economy really can’t handle any more of what we are doing.
C.O. is right about contact tracing, testing and quarantine, but politically, it would look like the Afghan surge seeing as the official line is we’ve already turned the corner and are on the road back. And we are, if we can just “Live and let Die”.
The question is one I’ve asked before: what level of compliance is necessary to have any effect? It won’t be the case that even the smallest amount of compliance will produce a measurable effect if for no other reason than because of network effects. 1%? 10%? 99%? If the level of compliance required to produce effects is high enough, it makes no sense.
” How do you go about tracing a person you happened to be sitting in the same car with on the subway? ”
In general you dont since only about 10% of people use public transportation even once a week. Tracing concentrates on the high risk encounters. We shouldn’t let perfect be the enemy of good. We may find tracing works better in some areas than others.
Steve
“what level of compliance is necessary to have any effect”
Let me do some rough figures. We will use the State of Ohio.
First we figure out size of the outbreak. Assuming an IFR of 0.5%, an infection period of 15 days, and an R of 1. They have 50 deaths per day, which implies 10000 people are currently infected. And 10000/15 = 667 people get sick every day.
Assume only half get tested, and only 10% of the tested are isolated – thats 33 people. But the isolation is 100% effective and they stop spreading the disease.
After 15 days, the number of active infections is about 9500. The R is lowered from 1 to 0.95. After 30 days, about 9000 cases vs 10000 cases if no isolation was in place.
If compliance is 50%. That is 167 people per day in isolation and after 15 days, the number of active infections is 7500. After 30 days, 5500 or so.
Between 10 and 50% compliance makes a noticeable difference within a month.
Since that has emphatically not been the observed results of the lockdowns it leads to one of two conclusions and likely both. Either compliance has been too low to be effective or the assumptions are wrong.