Azeem Azhar

March 15, 2020

This Is A Moment To Be Sanguine.

Read This Letter With That in Mind.

For readers outside of the UK, please note the parochial focus of this particular missive touches on many issues that might be relevant in your country. 

[T]he allocation criteria need to guarantee that those patients with the highest chance of therapeutic success will retain access to intensive care.

These are the new guidelines for the Italian College of Anesthesia, Analgesia, Resuscitation and Intensive Care, written an eon ago on Monday when Italy had 40% fewer cases of coronavirus than as I type these words. (There has been an update reported by the Daily Telegraph, “This is how it is in a war.”)

As COVID19 continues its exponential growth in most countries, the strategies which country select will have a massive impact on the severity outcome of this pandemic. Ignoring science at the top of government is unacceptable. However, science must not equally be a shield under which leaders absolve themselves of their political responsibility.

Writes my friend Nicolas Granatino, in a recent issue of his newsletter, Coronadaily reflecting on the British government’s approach to tackling the epidemic.

Boris Johnson and two senior science advisors delivered an impressive presentation. but their strategy for tackling the spread of the virus is controversial. As is becoming clear, the British government’s approach to handling the pandemic has been understood, in a nutshell, rightly or wrongly, as “let it roll through the population, avoiding tough containment measures because Brits will get bored, and protect the old until enough of us have herd immunity.”

It turns out it is more complicated than that. 

Adam Kucharski, the epidemic modeler, author of The Rules of Contagion, and a member of one of the expert groups that comprise SAGE which advises the government, makes it very clear here.

A lot of modellers around the world are working flat out to find best way to minimise impact on population and healthcare. A side effect may end up being herd immunity, but this is merely a consequence of a very tough option - albeit one that may help prevent another outbreak.

Ian Donald, a psychology professor with expertise in anti-microbial resistance points out:

The govt strategy on #Coronavirus is more refined than those used in other countries and potentially very effective. But it is also riskier and based on a number of assumptions. They need to be correct, and the measures they introduce need to work when they are supposed to…

This is probably the best strategy, but they should explain it more clearly. It relies on a lot of assumptions, so it would be good to know what they are - especially behavioural.  (emphasis mine.)

Models, even scientific ones, tell different stories, depending on the assumptions you put in. The assumptions are often uncertain and represent judgments & choices. The outputs of models have variances, based on the assumptions you make.

And an epidemiological model is a complex one with many inputs, variables, feedback loops and delays. (See this video here.)

These outputs become scenarios, which are based on your assumptions and judgments. Scenarios which you can navigate and use to explore future possibilities.

How you validate those assumptions is critical. How you select them is critical. How you make those judgments is also critical. It is also a human process that lends itself to argument and debate.

It is not sufficient to say “We are using the best science” and leave it at that. And certainly not in the era of open science, open data, and common tools to foster a healthy critical debate.

In the UK, the government gave us one choice, one approach, “from the science” as a fait-accompli. 

But no model works like that. Science doesn't work like that.

Behavioral scientists have challenged one tenet of the British government’s plan: that of behavioral fatigue. Here a stampede of behavior scientists write: (I don’t know what the collective noun for behavioral scientists is. Stampede seemed a reasonable choice)

we are not convinced that enough is known about “behavioural fatigue” or to what extent these insights apply to the current exceptional circumstances. Such evidence is necessary if we are to base a high-risk public health strategy on it.

An abelian of mathematicians has also written to challenge the UK government’s approach.  

Science doesn’t work the way Boris Johnson suggested it did. 

Especially not in the context of health decisions. Readers will remember part of my tooth fell out a couple of weeks ago. Even my dentist offered me the choice of yet another filling or a crown. Science drove my dentist’s assessment, and I was still offered two choices.

There are discussions, arguments, challenges, disagreements, external trade-offs that went into the final recommendations. But with millions of lives at stake and an entire economy built with and enabling those lives, there needs to be a deeper debate, deeper scrutiny.

The doctors in Lombardy have had to become utilitarians, applying clear criteria of survivability on the decisions they make. I support them--on the front line they have few luxuries.

But the British government is making similar decisions with many more variables, without clearly explaining the choices they faced and the roads they took. 

There is no doubt about the capabilities of the Chief Scientific Officer or Chief Medical Officer or the work of epidemiologists in various expert groups.  Or the quality of their models. (Although no model is perfect, many are wrong and some may even have catastrophic mistakes in them. Think back to the models driving financial risk in the run-up to the global financial crisis. Remember David Viniar’s bullshit about “twenty-five standard deviation” events?)

It is that the modeling only throws light on a couple of aspects of the problem at hand. The model’s job is to also outline the possible paths of specific political choices. 

It was a curious approach by the British government. One which has lost the trust. Yet, as I argued two weeks ago, public health emergencies require public trust.

the current British government, fought its electoral campaign with explicit attacks on the mechanisms of the state that stand outside politics, such as the judiciary and the Civil Service. At a moment when trust is most needed, to manage this burgeoning public health crisis, the government has been involved in a vicious attack on the very institutions that engender that trust.

How do you regain trust?

You can brief journalists and you can get your more articulate ministers to write an article in friendly newspapers. This is an industrial-age approach that might have worked with the poorly-educated, information-starved, under-networked, populace of the 1930s.

In the exponential age, your population is vastly more educated, resource-rich and networked. It has access to pretty much the same international research you do. It has, in many cases, better capabilities than the government can rely on.

And we can look on the Internet and read what is happening in South Korea, Taiwan, Singapore, Spain, Italy, Norway, Iceland, the Faroe Islands, China and you name it. We can see what they are doing. The temperature guns. The drive-through testing.

Italy looks like it can flatten the curve.

South Korea (below) is squishing new cases even as it tests vigorously. We can see this because the information is available from reliable sources.

We can hear Andy Slavitt, Obama’s healthcare majordomo, sound the alarm for the impending tsunami.

Nearly 30m of us have read Tomas Pueyo’s “Coronavirus: Why you must act now”. Some will have read the critique of whether “flattening the curve is a delusion.” You might have seen the Washington Post’s visualisations on how flattening the curve works.

Or perhaps you have been following the work of Yanner Bar-Yam and colleagues at the New England Complexity Science Institute whose models suggest “massive testing can stop the coronavirus outbreak” or their earlier work on Ebola contact tracing which has led them to suggest that even low compliance towards travel can stop outbreaks.

Many will wonder how quickly we could administer drugs to stop the cytokine storm that kills many who die of Covid-19. Or could we expand testing of and access to remdesivir, chloroquine or viagra, all of which seem to have some efficacy.

Some will ask whether acting aggressively now to slow the spread could buy time for health capacity to build up, even marginally. Or for Senegal to finish work on its 10-minute PCR test kit, or for us to validate Singapore’s serological tests.

In 2002, most Brits do not sit around doing the Daily Telegraph crossword, sipping tea, eating crumpets, waiting to lap-up the latest announcement from Downing Street on the wireless. So it’s truly bizarre to present a plan (perhaps it is the best??) that is such a massive outlier to the rest of the world and not expect pushback.

Openness, transparency about these types of decisions is a pillar of a democratic society. And it is foolish to think you can get away with a decision like this without someone asking difficult questions. Much more so when scientists are willing to state they disagree with you. And even more so in the era of the Internet. ( Paul Ginsparg demonstrated the value of open access when he launched Arxiv in 1991. Arguably Gutenberg did if even earlier.)

A better approach would be to open the model, to open the assumptions, for constructive discussion in the brief moments we have available, in a way that could catalyze improvements. You still have to lead, as Prime Minister, without question but you have to bring people with you. And we only at most a few weeks before the tsunami hits and our hospitals are, likely, overwhelmed.

Playing with one epidemic simulator, as it happens, helped me better understand Sir Patrick Vallance and Chris Witty, the UK government’s advisors rationale. Don’t get hung up on the model simulation I used. It is a toy simulator. Its conclusions are mostly useless. I don’t know what was under the hood. And it almost certainly ignores many of the more nuanced variables and relationships the real models used by SAGE have.

This toy simulator irrelevant except that:

• It is clarifying for a citizen to play with a model like this, fiddle with assumptions, run simulations. It helped you understand the trade-offs that we face in tackling this epidemic—and it starts to outline what our personal responsibility and capacities may be.

• The trade-offs are complex with non-linear effects of such a baffling nature I can only describe them as counter-intuitive. (Which more shows up the weaknesses in my expertise, more than anything else.)

I have no deep professional expertise on whether these epidemiological models. But I do understand the useful limits of models and can grok that the model is one input into an overall decision. That there are nuances. I do know there are assumptions on which decisions are made. The British government failed to give us their assumptions, the tools and democratic rigor to allow us to come to any kind of conclusion.

They still have a chance to do that. And early signs are that the British Government is demonstrating some pretty creative thinking (especially in enlisting the private sector and ramping up the NHS, our vehicle for universal, socialised medicine.)

But in being clearer and more inclusive in their communications, they might enlist creativity and talent and trust around the country in myriad other ways—not least in our willingness to step up and do the right thing.

What if we are more resilient and self-organizing and willing to clamp down on contact immediately? And will we do, as my community is, to support the vulnerable at a community level? What if we, the people, understand the horrific consequences of widespread promiscuous lunching at the local boozer?

Public health crises require public trust. A crisis like this, as we are learning from Italy, is more like a war. And it will be a long, grinding war, for which citizen trust, forbearance, and participation will be required.

As I went to bed on Saturday night, #boristhebutcher was, perhaps unfairly, trending on the socials. Such is the cost of forgetting about open, transparent, interactive communication that the internet enables.

This is the shape of democracy in the exponential age.

My friend from Shenzhen, six weeks into quarantine, sent me a WhatsApp. I know he won’t mind me sharing it with you: “Covid19 cases going up in Europe. Pls, take care.”

And let me add, keep your distance and WASH YOUR HANDS.

 Azeem

P.S. On the point about the economy and its importance during a crisis… we’ll need a healthy economy of sorts to survive this marathon. The economy provides the goods and services that keep our population fed, nourished and healthy, and creates capacity in our health system. This isn’t so much an argument about GDP growth, but more than a healthy economy enables a healthy population and a functioning medical system.

Thanks to Nicolas Granatino, Vishal Gulati, Hampus Jakobsson, Paola Bonomo and many others for discussions that helped with this letter.

Thanks for subscribing to Exponential View.

This post is public, so feel free to share it. Share