Rotundus.com

Class, Race, Coronavirus, And Cuisine

To solve a problem you need to ensure that you accurately understand it. How much of the increased death rate amongst African Americans that's been highlighted on the news is due to poor diet, and so what extent is the diet in question not the food of African-Americans but rather the food of the South (including white Southerners). If these figures are accurate might the death rate be disproportionately white in some states as well?

Corruption information and vote share: A meta-analysis and lessons for experimental design

"Debate persists on whether voters hold politicians accountable for corruption. Numerous experiments have examined if informing voters about corrupt acts of politicians decreases their vote share. Meta-analysis demonstrates that corrupt candidates are punished by zero percentage points across field experiments, but approximately 32 points in survey experiments."

Gender-based homophily in collaborations across a heterogeneous scholarly landscape

To quote Matt Grossman's summary tweet: "Researchers are more likely to have same gender co-authors in many scientific fields (jstor data), creating segregated networks; gender homophily is higher with more women’s representation because women have more women co-authors available".

It's interesting to compare Vox's take on whether or not the current coronavirus of concern escaped from a lab to their takes on lab safety in roughly the past year. This for example is what Vox published in March 2019:

Thread on biosecurity and pandemic pathogens: So, sometimes something goes terribly wrong at a high-security lab and a pathogen escapes. This is not, actually, itself a sign we're doing something wrong. (1/?) https://t.co/51rQR3AUc2

— Kelsey Piper (@KelseyTuoc) March 20, 2019

They then in September published another article noting that a Russian labs which contained samples of smallpox and ebola exploded.

This compares to a March 2020 article with the following subheading:

There’s a rumor the coronavirus started in a Chinese lab. And a scientific consensus it didn’t.

The article then elaborates on what this might mean:

In one version of the rumor, the virus was engineered in the lab by humans as a bioweapon. In another version, the virus was being studied in the lab (after being isolated from animals) and then “escaped” or “leaked” because of poor safety protocol.

The first version of this seems likely to be false based on various evidence that I've seen reported - i.e. it seems highly unlikely this is an engineered bioweapon. The second though seems somewhat more probable to me. What's the defense the article offers against the second explanation?

These are the two references to experts that the article cites:

• Jim LeDuc, head of the Galveston National Laboratory, a level 4 biosafety lab in Texas: “I can tell you that lab in Wuhan is equivalent to any lab here in the US and Europe"

• Gerald Keusch, a professor of medicine and international health and associate director of Boston University’s National Emerging Infectious Diseases Laboratories: "I don’t think there’s any likelihood that the lab is less prepared in terms of protocol and capability than any lab in the US. It’s really good, though nothing is perfect"

It seems as though that lab in Wuhan was likely about as safe as comparable labs in the US and Europe, but if you compare to Vox's article from 2019 embarrassing failures of lab safety procedures have also occurred in both regions.

I wonder if the best way to look at the situation re: safety is in that quote from Keusch on safety there: "really good, though nothing is perfect".

The whole situation makes me think of another sort of thing that a lot of safety precautions are taken regarding: nuclear weapons. Read Command and Control: Nuclear Weapons, the Damascus Accident, and the Illusion of Safety, and I suspect you might walk away thinking that, in spite of the precautions taken, that it's almost surprising that no nuclear weapons have been accidentally detonated to this point.

There's no definitive proof, but for me The Trail Leading Back to the Wuhan Labs was enough for me to update my best guess as to the source of the current pandemic to being a lab containment failure.

Despite this, I don't think that it's reasonable to blame China for being the site of such a containment failure if indeed one happened there as they seemed to be taking precautions that met standards of comparable labs around the world. Roll the dice often enough at labs around the world and eventually one will probably roll a bad result.

I do think that it is reasonable to blame China for its efforts to suppress information about this outbreak (which I definitely think it was doing, given treatment of original whistleblowers and whatnot, regardless of whether or not this was indeed a lab containment failure). I suspect that given the embarrassment if would cause the host country of the lab to have a pathogen escape would like cause many countries to follow a similar strategy of suppression - and again I want to stress that it doesn't appear that the labs in question were operating with lower safety standards than their counterparts around the globe.

Some might say that the current pandemic means that dangerous infectious diseases shouldn't be studied in labs, but I tend to think they should be. It does leaves researchers investigating these diseases in somewhat of a Catch-22 situation though - i.e. you're stuck trading off between a slightly increased chance of outbreaks with an increased chance of finding better ways to fight outbreaks (which I think are inevitable whether or not these diseases are studied in labs).

There are a few angles that have me thinking back to the history of medication and the sort of lessons that might be drawn from it.

Drug discovery

I've mentioned before that what science has looked like throughout history often differs from many in the scientific community seem to think is has. Major journals like The Lancet and Nature only started sending out papers for external peer review in the 1970s for example. The same seems to apply to how some critical drugs were developed:

Large effect size drugs like penicillin and insulin were not discovered via double-blind randomized controlled trials.

They were found via an iterative process. Call it results-based medicine.

Penicillin: https://t.co/bknTjboKmW

Insulin: https://t.co/PGMiFQg1aO pic.twitter.com/gVnNt1bkfF

— Balaji S. Srinivasan (@balajis) March 25, 2020

Fighting pandemics with mandatory masks

So get this. In the Spanish flu, San Francisco authorities mandated all people wear masks in public. Police arrested citizens who wouldn’t wear them. For a whole month this pretty much worked.

But then

— T. Greer (@Scholars_Stage) March 23, 2020

One problem noted there - and you might expect with current measures - are questions of just how long the public is likely to be willing to tolerate them.

Variolation

You may not have heard much about this but it's the precursor to modern vaccinations. It involved deliberately infecting people with a small dose of smallpox in order to significantly improve survivability from the infection. I'd heard of this before, but in the covid19 context it was Robin Hanson that seems one of its early vocal advocates as a possible way to address the current pandemic.

The medical literature consistently finds strong relations, in both animals and humans, between initial virus dose and symptom severity, including death. The most directly relevant data is on SARS and measles, where natural differences in doses were associated with factors of 3 and 14 in death rates, and in smallpox, where in the 1700s low “variolation” doses given on purpose cut death rates by a factor of 10 to 30. For example, variolation saved George Washington’s troops at Valley Forge.

He notes that using lockdowns to try to prevent of a disease may have the effect of increasing the fatality rate from the a disease like the one behind the current pandemic, as being trapped in close confinement could cause an initially infected person spreading the infection to those they live with by in higher (and thus more dangerous) initial doses.

Now I see Scott Gottlieb talking of viral load, pointing to an New York Times article. The author of that article spends much of it talking about the risk which is certainly a legitimate concern. He says this:

It would be unethical to experimentally manipulate viral dose in humans for a pathogen as serious as the coronavirus, but there is evidence that dose also matters for the human coronavirus. During the 2003 SARS coronavirus outbreak in Hong Kong, for instance, one patient infected many others living in the same complex of apartment buildings, resulting in 19 dead.

I keep finding myself thinking to what extent I should buy the arguments of the current medical ethicists though. At the moment it seems that one major reason why the US has failed to respond well to the current pandemic seems to have been concerns by the CDC and the FDA over ethics and safety regulations¹.

Here I think Hanson's words are useful - I suggest reading the whole post. It's worth knowing thatthere's not yet adequate data to answer some questions and testing which does legitimately pose a risk would be needed to address them. Controlled infection is one strange way to go about actually "flattening the curve", could help move society closer to herd immunity, and could have savings in lives that might overwhelm any lost.

Of course, if low doses cut death rates by a factor of two or more, variolation volunteers would actually cut their chance of death, perhaps greatly. Yes, the first few thousand volunteers could be less sure of such gains, but they could be compensated for this risk, just as we now consider compensating subjects in vaccine trials using live Covid19 viruses. We could pay variolation volunteers cash, offer their loved ones priority medical care, certify them as safe for work and social gatherings, and honor them like soldiers selected for their elite features who take risks to produce community gains.

It seems to me that often people look at risk as something you either take or you don't rather than looking enough at offsetting one type of risk against another.

Interestingly, since the last post here it seems that both the CDC and US Surgeon General have doubled down on advice to avoid use of face masks. Still suspecting that that advice may be changing soon regardless, given other moves. Here's another tweak from Scott Gottlieb the former FDA commissioner for example:

Useful primer on the benefits of masks. A recommendation from CDC for consumers to wear cotton masks (with guidelines on how to fashion these products from household items) can improve safety and reduce spread and not strain the hospital supply chain. https://t.co/yAEi8hcLAH

— Scott Gottlieb, MD (@ScottGottliebMD) March 30, 2020

One reason that I've concentrated on him is that he seems to be one of the few both having a history of being critical of the Trump administration's response while, particularly lately, still being respected by those in the administration. Encouraging the public to wear masks is also a component of the report he's put together showing how to reopen society. Speaking to Good Morning America, Dr. Fauci - a face you're likely to see in Trump press conferences - also seems to find some support for the position.

At the moment, looking at the still emerging research, it seems as though the Czech Republic - the first country in Europe to have required wearing masks in public - seems to have successfully flattened the curve. Just within the last day or so the Germany Health Ministry is starting to echo the advice of the German Medical Association to wear masks, Austria has mandated wearing masks in supermarkets and Bulgaria has made wearing masks in public places compulsory.

The Czech Prime Minster has been publicly calling for Trump to require masking in public and it seems to me that there are actually quite politically compelling reasons for Trump to do so. In general in politics one need not tell a true story - it only needs to be true enough to be able to sell it to the public. If you look at how the CDC and the FDA obstructing testing are key reasons as to why the US is in the poor position it currently is, it becomes easier to sell a (generally false) narrative of Trump bravely fighting the Deep State. It doesn't matter that the Surgeon General has attempted to suck to Trump in very cringey and false ways - I look at Trump as a narcissistic sociopath who has no loyalty to anyone other than himself. If he can act in ways which seem to be ahead of organizations like the CDC and the WHO¹ in adopting policies that are later more generally accepted (and for which a decent base of evidence appears to continue to be building) that seems to provide a fair bit of cover for his prior (and continuing) incompetence.