The Roman Philosopher Lucius Anneaus Seneca (4 BCE-65 CE) was perhaps the first to note the universal trend that growth is slow but ruin is rapid. I call this tendency the "Seneca Effect."
Showing posts with label flatten the curve. Show all posts
Showing posts with label flatten the curve. Show all posts

Sunday, April 2, 2023

"Flattening the Curve." The Origins of a Bad Idea


In 2003, the "Anthrax scare" led many people to use duct tape to seal the windows of their homes to protect themselves from the deadly germs. It would have been a good idea if the purpose was to suffer even more than usual from indoor pollution while at the same time doing little or nothing against a hypothetical biological attack. Yet, this folly was recommended and encouraged by the national government and local ones. It was the first taste of more to come. I already mentioned this story in a previous post, but here I'll go more in-depth into the matter. 


We are still reeling from three years of madness, but it seems that many of us are starting to make a serious effort to try to understand what happened to us and why. How could it be that the reaction to the Covid epidemic involved a set of "measures" of dubious effectiveness, from national lockdowns to universal masking, that had never been tried before in humankind's history? 

For everything that happens, there is a reason for it to happen, and the "non-pharmaceutical interventions" (NPIs) that were adopted in 2020 have their reasons, too. Some people speak of global conspiracies and some even of evil deities, but the origin of the whole story may be more prosaic. It can be found in the development of modern genetic manipulation technologies; a new branch of science that started being noticed in the 1990s. As it normally happens, scientific advances have military consequences; genetic manipulation was not an exception. 

Not that "bioweapons" are anything new. The chronicles of ancient warfare report how infected carcasses of animals were thrown inside the walls of besieged cities and other similar niceties. More recently, during the 19th century, it is reported that blankets infected with smallpox were distributed to Native Americans by British government officials. Overall, though, biological warfare never was very effective, and not even smallpox-infected blankets seem to have been able to kill a significant number of Natives. Besides, biological weapons suffer from a basic shortcoming: how can you damage your enemies only while sparing your population? Because of this problem, the history of biological warfare does not include cases where bioweapons were used on a large scale, at least so far. The low effectiveness of bioweapons is the probable reason why it was not difficult to craft an international agreement on banning their use ("BWC", biological weapon convention), ratified in 1975. 

Up to recent times, bioweapons were not seen as much more dangerous than normal germs, and the generally accepted view on how to face epidemics favored a soft approach: letting the virus run in the population with the objective of reaching the natural "herd immunity." For instance, in a 2007 paper, four respected experts in epidemiology still rejected such ideas as confinement, travel bans, distancing, and others. On quarantines, they stated that "There are no historical observations or scientific studies that support the confinement by quarantine of groups of possibly infected people for extended periods in order to slow the spread of influenza." But military planners were working on the idea that bioweapons would be orders of magnitude deadlier than the seasonal flu, and that changed everything. 

Genetic engineering technologies were said to be able to create new, enhanced germs, an approach known as "gain of function." Really nasty ideas could also become possible, such as "tailoring" a virus to attack only a specific ethnic group. Even without this feature, a country or a terroristic group could develop a vaccine against the germs they created and, in this way, inflict enormous damage on an enemy while protecting the country's population (or the select few of an elite aiming at global depopulation). 

Fortunately, none of these ideas turned out to be feasible. Or, at least, there is no evidence that they could be put into practice. That doesn't mean they were not explored, even though research on biological weapons is prohibited by the BWC convention. Researching new germs is orders of magnitude less expensive than making nuclear weapons, and so it is also feasible for relatively poor governments. (biological warfare is said to be the weapon of the poor). Whether truly effective bioweapons can actually exist is at least doubtful, but supposing that they could exist, then it makes sense to prepare for a possible attack. 

The "anthrax scare" of 2001 was the first example of what a modern terror attack using bioweapons could look like, even though it was nothing like a weapon of mass destruction, and nobody to this date can say who spread the germs using mail envelopes. Nevertheless, it was taken seriously by the authorities, and it led to the "bioterrorism act" in 2002. Shortly afterward, Iraq was accused of having been developing biological weapons of mass destruction, and you may remember how, in 2003, Colin Powell, then US secretary of state, showed on TV a vial of baby powder, saying that it could have been a biological weapon. Over the years, several government agencies became involved in planning against bioweapon attacks, and the military approach to biological warfare gradually superseded the traditional plans on how to deal with ordinary epidemics. 

The idea was that a truly deadly virus attack would cripple the infrastructure of a country and cause immense damage before the germs could be stopped by a specifically developed vaccine. So, it was imperative to react swiftly and decisively with non-pharmaceutical interventions ("NPIs") or "measures" to stop the epidemic or at least slow it down. This idea was rarely expressed explicitly in public documents, but it was clearly the inspiration for several studies that examined the effects of an abnormally deadly virus. One was prepared by the Department of Homeland Security in 2006. Another one comes from the Rockefeller Foundation in 2010, where you can read of a scenario called "Operation Lockstep" that described something very similar to what came to pass in 2020 in terms of restrictions. 

These military-oriented studies were mostly qualitative. They were based on typical military ideas such as color-coded emergencies, say, "red alert," "orange alert," and the like. For each color, there were a series of recommended measures, but little was said about why exactly certain colors were chosen to be coupled with certain actions. But there were also attempts to quantify the effect of NPIs. A paper on this subject was published in 2006 by the group of Neil Ferguson of the Imperial College in London, where the authors endeavored the task of disentangling the effects of different measures on the spreading of a pathogen. Factors such as home quarantines, school closures, border closings, reduced mobility, and others were examined (interestingly, the widespread use of face masks was not considered). The study didn't expend much of an effort to compare guesswork with real-world data, but it was not so bad in comparison with the harmless mumbo-jumbo that scientists normally publish. Let's say that it might have been an interesting exercise in epidemiological modeling but nothing more. The problem was that it came at a moment in which "biological warfare" was all the rage and that it may have influenced later military planning. 

A study that had an enormous influence on the (mis)management of the 2020 pandemic was directly inspired by Ferguson's paper. It was proposed as a CDC report by Rajeev Venkayya in 2007, who presented his model in the form of the "double curve" that later became famous. Here it is; it is the origin of the "flatten the curve" meme that became popular 13 years later. 



Remarkably, Venkaya's model was completely qualitative. The curves were just a rendition of those proposed by Ferguson et al. without any attempt of quantification. Venkayya's paper didn't attain great popularity, and it remained dormant for more than one decade until the Covid-19 pandemic arrived. Then, suddenly, the double curve became highly fashionable. It went viral and literally "exploded" on the Web and in the media.  

During the early days of the Covid-19 epidemic, one of the main supporters of the double curve model was Tomas Pueyo, who was perhaps the first to use the term "flattening the curve." His post of March 10, 2020, on the matter had more than 40 million visualizations. Notably, Pueyo's previous expertise was in engineering and communication, and he was known for a discussion about the characters of the "Star Wars" series as role models in business. So, you might be reasonably perplexed about the authority that he claimed to have on epidemiology. Nor is it clear who pushed his blog to the first places of the search engines. In any case, Pueyo showed no signs of undue modesty in the way he considered himself an expert. For instance, he wrote: 

As a politician, community leader or business leader, you have the power and the responsibility to prevent this.

You might have fears today: What if I overreact? Will people laugh at me? Will they be angry at me? Will I look stupid? Won’t it be better to wait for others to take steps first? Will I hurt the economy too much?

But in 2–4 weeks, when the entire world is in lockdown, when the few precious days of social distancing you will have enabled will have saved lives, people won’t criticize you anymore: They will thank you for making the right decision.

Pueyo was even invited in TV, where he had a chance to vocally disparage the concept of "herd immunity," which has been a staple concept in epidemiology for at least a century. If you listen to him in that TV show, you'll probably notice that he provided no evidence that he understood how epidemics spread in a population, but that's not surprising for someone whose expertise is mainly in marketing and communication. Look at minute 12:30 of the TV interview to see Pueyo's face and gesture when the scientist being interviewed mentions herd immunity. Clearly, Pueyo had no idea of what herd immunity is. It is amazing that this bizarre character was allowed to have so much influence on global policies. But so it went, and the double curve was accepted as scientific wisdom and the undisputable target for governmental interventions everywhere. 

I discussed the shortcomings of the double curve model in a previous post. Basically, the model was flawed because it didn't include methods to verify whether the measures were doing something or not. So, the results of the attempt to "flatten the curve" were modest, if they existed at all. But the story didn't end there. There does not exist something so bad that someone can't make it worse.

Little more than a week after launching his "flatten the curve" post, on March 19, 2020, Tomas Pueyo was on the march again, and he went rapidly forward in the direction of the worse. He got rid of the only thing that had some contact with reality in Venkayya's model: the fact that epidemic curves are bell-shaped. Now Pueyo said that the natural shape of epidemic curves is exponential growth and that only specific measures could force it to bend down into a bell-shaped curve. That was probably a necessary consequence of the fact that Pueyo never really understood the basics of epidemiology. So, having started with the wrong assumption (exponential growth), he moved to the wrong conclusions. Starting with an unrealistically large value for the fatality rate (3.4%), he proposed the graph below, defined as "The Hammer and the Dance." The idea was that the epidemic curve would grow to infinity unless it was brought down by means of harsh and immediate measures of containment (the hammer) and then by keeping it low with lighter measures (the dance). The data showing that this could be done..... data? What data? 






This diagram, evidently, doesn't come from an epidemiological model, not even a highly simplified one. It is just drawn accordingly to what Pueyo thought the behavior of the system should be. 

The story of the hammer and the dance was the start of a disastrous debate (let's call it this way) where plenty of people came up claiming that herd immunity was a flawed concept and that the only way to avoid being all infected, and many killed, was to enable containment measures. It also led to the concept that the curve could not just be flattened but "crushed." Again, we can see the influence of the military approach: terms such as "flattened" or "crushed" are typical of warfare, but viruses cannot be killed using military weapons. In any case, the concept of "crushing the curve" gave rise to the idea of "Covid Zero." It was another disaster that befell us, sending entire countries to a human and economic disaster in the desperate search for an unattainable goal. 

Mostly, the whole disaster was due to our inability to understand how models work. Formal mathematical models are a recent feature of the human way of dealing with reality. They are supposed to help you understand how the world works and even predict how it will evolve. But you have to be careful: the model is not the reality, just like a map is not the territory. A wrong model is not necessarily dangerous, but it can be. A military model that tells you that attacking Russia in winter is a good idea is a good example. The idea of "crushing the curve" is another example. 

Will we ever learn how to use models? Maybe. But, for the time being, models play the role of guns handled by children.