A systematic error that lead to a bad policy response to COVID-19

by ChristianKl4 min read12th Jun 20217 comments

23

Covid-19World Optimization
Frontpage

Introduction

When it comes to dealing with adopting good COVID-19 policies, our civilization failed. This suggests low ability to respond well to another future crisis that might be an X-risk. While we made multiple classes of errors this post will look at one particular class of errors. The failure to come to synthesis through dialectic reasoning. In dialectics, the Fichtean "thesis–antithesis–synthesis" model,  when there are two opposing positions A and B there’s the idea that it's often possible to develop a new position S that builds on the ideas of both positions with S being superior to both A and B.

Examples

Requiring FFP-2 masks

Thesis: 

FFP-2 masks should be required because it’s what medical professionals use to protect themselves against infection.

Antithesis: 

Medical professionals get regular fit-testing while only 25–75% of people without fit-testing use them successfully and most of the benefit they provide over OP-masks comes from using them correctly.

Synthesis: 

Require FFP-2-masks and make fit-testing easily available for non-medical personnel. 

Create an additional FFP-X category, where the manufacturer only needs to do one study to get them to the market and that study consists of giving a representative group fit-tests with the mask. Require the seller of the masks to tell the customers about the fit-test success rate of the mask

Allow DIY antibody tests

Thesis: 

DIY antibody tests allow people to reduce the chance to infect other people by staying home when the test is positive.

Antithesis: 

The high false-negative rate of DIY antibody tests gives people false confidence and potentially induces them to reckless behavior.

Synthesis: 

Easy: Allow DIY antibody tests but require publishing sensivity/specificity in the product name whenever the product is sold or otherwise marketed.

Hard: Create a central app that gathers as much information as possible and that allows interpreting the test for the user given his background variables and that calculates MICRO-COVID for infecting other people for user selected activities.

The Situation in Berlin

In both examples we had a policy debate where the authorities first argued the antithesis  and after the thesis good enough support switched to and not the synthesis. 

This is surprising because the synthesis would provide more benefits, coming up with the synthesis isn’t very hard, politicians listen to credentialed experts and there are high benefits to getting it right both for their reelection chances and for the good of society due to less deaths, disability and economic damage.

The Fichtean "thesis–antithesis–synthesis" found little discussion on LessWrong because it seems too simple and trivial to be useful. The current situation falsifies the hypothesis that the process happens on its own without working to make it happen.

My hypothesis for why the error happens

Given that the question has very high personal utility for everyone and impacts everyone personally, people attach a lot of value to getting the question right. When everyone attaches a lot of value to getting a question right, the question becomes political. 

In political domains individuals gain status if they advocate the same position as their allies. Advocating the thesis instead of advocating the synthesis is a better tribal signal.

It’s a bad move for a political actor to create a synthesis themselves when there’s a lot of status in advocating either the thesis or antithesis because allies advocate it.

In contrast there’s the strategy of being a leader. A leader can propose a synthesis and win status by convincing others but this move carries a lot more responsibility and thus risk of losing local status then going along with the position of allies. 

Politicians proclaim that they listen to the scientists as a move to avoid responsibility and pass the ball. Unfortunately, the layer a step down is still very political and afraid of thinking for themselves and doing synthesis. 

When people with the intellectual ability at Harvard think up a Covid-Roadmap the resulting document is devoid of synthesis. This suggests that the level at which the document is created is too political for synthesis. 

The dynamics are similar to that of a moral maze which multiple layers of political management produce a dysfunctional organization.

What can we do as rationalists?

Writing posts about individual clashes

According to Cummings, our community managed to speed up the first lockdown in the United Kingdom by writing blog posts that were read by multiple people in Number 10. While this is partly due to Cummings being a fellow rationalist and thus more inclined to listen to us, well written posts can be easily passed around. 

A political actor that advocades a thesis doesn’t lose the tribal benefits of advocating the thesis when he passes around a link to another person advocating a synthesis. If a post gets passed around enough, it creates public knowledge of the synthesis having support and it becomes safe for political actors to advocate it. 

Understanding the reasoning failure better

While I proposed a model, there might be better models. Part of the rationalist project is improving general decision making and we should produce models and propose general solutions.

Conclusion

Any large crisis is going to produce the same dynamics where actions regarding the crisis become highly political. Given our political structures with multiple layers of political actors at power positions, we will get bad policy responses to the next crisis as well unless we act and change our structures or create new structures.

23

6 comments, sorted by Highlighting new comments since Today at 7:59 PM
New Comment

Mask synthesis: Use elastomeric respirators. Elastomerics offer better fit and more protection (N100) than any disposable PPE. If necessary, develop respirators that fit even better with little to no fit testing (like PAPRs).

Expert consensus: "Don't wear a mask...oops, wear a cloth or surgical mask until you get vaccinated." This advice has led to millions of preventable deaths and counting.

Conclusion: Don't worship experts. When making critical decisions, quadruple-check any expert advice (especially if it's about a soft science). If the advise doesn't make sense, disregard it and make a decision based on your own research and reasoning.

Mask synthesis: Use elastomeric respirators. Elastomerics offer better fit and more protection (N100) than any disposable PPE. If necessary, develop respirators that fit even better with little to no fit testing (like PAPRs).

I think you get to that position by using first-principle thinking which is a different way to reason then the dialetic way. Practically, first-principle thinking is also seldomly done by those making health policy but we should keep different reasoning strategies apart as rationalists when we want to understand how to think about thinking. 

As far as the substance matter goes:

"Use elastomeric respirators" is a decent personal decision if you are in a jurisdiction that doesn't require you to wear a FFP-2. It's not a general policy position. 

"Require everyone to wear elastomeric respirators" would be a policy position but there are certainly contexts where those are unpractical. 

One problem with elastomeric respirators is that they are generally designed for a use-case where filtering exhaled air isn't central for that use-case they often have ventils that allow exhaling unfiltered air. From the outside it's hard to know whether someone wearing a elastomeric respirators is filtering their exhaled air or isn't which makes it harder to enforce policies around them and many people won't understand that they should not use the exhaling vents. While it might be better policy it isn't a slam dunk. 

I understand that you're trying to analyze the policy failures in terms of dialectic reasoning, but the policy about masks that results from that reasoning is not good enough. So, perhaps first-principles thinking should be emphasized as a better way to formulate policy, at least in certain situations.

The policy position about masks would be to recommend the use of elastomerics (assuming adequate supply) but not to mandate their use. Mandates would only be required if compliance was so low that there was a high risk of hospitals becoming overwhelmed.

Assuming adequate supply of elastomerics again, vents are mostly a non-issue. If some people would want to risk infection by not wear an elastomeric, that would remain a personal decision. Also, the users of elastomerics would have a much lower chance of being contagious. Vents would be a problem only in special circumstances like nursing home care where disposable PPE would be more appropriate. [Ventless elastomerics are available such as the MSA Advantage 290, so disposable respirators seem completely unnecessary in any circumstance.]

There are certainly other possible ways to make address the mask issue.

I can improve my decision making over that of people who simply follow the official recommendation a large percentage of the population likely won't be able. In a situation like a pandemic I'm very dependent on the actions of others around me.

PAPRs have the potential to be better than elastomeric respirators. They can be cheaper (the DIY kind), filter incoming and outgoing air as well as an N100 filter, and fit testing and seal checks aren't needed.

https://www.viralhelmets.com

One more primary failure: Containment.
Thesis: Preventing travel (get the planes out of the air) from infected areas would contain the virus. Individual countries (including China), showed that virus could be controlled eliminated, if infections were not imported.

Antithesis: It costs too much, kills the tourist/travel industry; I want to travel; Restrictions on freedom are evil..

Well once it took hold, the restrictions happened anyway. Doing it early would dramatically reduced cost. If the symptoms had been as alarming as say Ebola, then I think early action would have been easier. "Coronavirus = cold" was all too common - maybe it should have been called SARS immediately?