Saturday, April 25, 2020

When and How Did the Experts Decide that the COVID-19 Pandemic Justified a Lockdown?

When and how did public health experts decide that the COVID-19 pandemic justified a lockdown of our social and economic life, for the first time in human history, even at the cost of a global great depression?

For the past six weeks, I have tried and failed to answer that question.  If anyone has the answer, I would be pleased to hear it.

The political leaders who have ordered lockdowns--Trump, most of the state governors in the U.S., and other political leaders around the world--have said that they are following the recommendations of epidemiologists and other public health experts.  This implies that the experts have weighed the social and economic costs of a lockdown that creates a global depression against the public health benefits of such a lockdown, and they have demonstrated that the benefits clearly outweigh the costs.  Strangely, however, at least as far as I can tell, no one has actually carried out such a cost-benefit analysis that would support the moral judgment that a lockdown is justified as serving the public good.

They invented the term "social distancing" and formulated the "6 feet rule."  They developed the public health policies for "flattening the curve" of infections in an infectious disease pandemic to minimize deaths until a vaccine could be developed.  In 2006, doctors and public health experts working for the U.S. federal government conceived these ideas as part of a strategy for using "Non-Pharmaceutical Interventions" (NFIs) to combat a deadly pandemic.  In 2007, their proposals were officially adopted by the U.S. government as guidelines for mitigating a future pandemic of infectious disease.  In 2017, in the first year of Trump's administration, a slightly revised version of these guidelines were renewed as official government policy.  Then, in the first three months of 2020, some of the original proponents of these guidelines recognized the novel coronavirus coming out of China as creating the pandemic they had been preparing for, so now was the time for the U.S. government to implement the guidelines they had designed in 2006.  Federal bureaucrats in the CDC, the NIH, and the Homeland Security Department were then able to persuade President Trump and most of the state governors to enforce these guidelines, which led to the lockdown orders.

Some of this story has been told in the New York Times (here and here), which includes a series of email exchanges from the beginning of this year (here).

At first, I thought I could find the answer to my question here.  But after studying the pertinent documents, I still cannot see when and how the public health experts demonstrated to themselves and others that a moral cost-benefit analysis showed that the costs of a global depression caused by a governmental lockdown of social and economic life were far less than the public health benefits of this policy.  This makes me wonder whether political leaders thoughtlessly stumbled into issuing lockdown orders without anyone having done the moral analysis that would rationally justify this.

It all begins with a speech by President George W. Bush at the National Institutes of Health on November 1, 2005.  Bush had read John Barry's newly published book on the flu pandemic of 1918--The Great Influenza.  From reading this book and talking with some public health professionals, he worried that the world was unprepared for another such deadly pandemic.  So he proposed that Congress should appropriate $7.1 billion in emergency funding to support a comprehensive national strategy to protect the American people from a future pandemic.

His strategy was to meet three goals--an international system for detecting pandemic outbreaks early, developing and stockpiling vaccines and antiviral drugs, and having plans at the federal, state, and local levels for responding to any pandemic that reaches America.

To develop the plans for responding to a pandemic, the Bush administration asked Richard Hatchett, a White House advisor, and Carter Mecher, a VA physician, to organize this.  They were joined by Howard Markel, director of the University of Michigan's Center for the History of Medicine, who had been working with a Pentagon research group studying the history of the governmental responses to the 1918 pandemic.

Their interpretation of what was done in the 1918 pandemic was crucial for what they proposed, because they saw the relative success of the city government in St. Louis in reducing deaths from the flu through quick interventions to slow the spread of the virus--in contrast to the slow response of the government in Philadelphia--as setting the standard for "non-pharmaceutical interventions" against pandemics in the future.

On October 8, 1918, in response to the growing number of flu cases, the city government of St. Louis announced the closing of theaters, movie houses, pool halls, schools, churches, and all large public gatherings.  Beginning on November 13, this closure order was gradually lifted.  But then, two weeks later, on November 27, there was a new closing of the schools and a ban on public gatherings, because there had been an increase in the flu infection rate.  This second closure was lifted at the end of December.

Notice that neither of these two closure orders required any general lockdown of the social and economic life of St. Louis.  Retail stores, businesses of various kinds, and factories remained in operation.

There was one order in St. Louis closing all "non-essential" stores, businesses, and factories; but it lasted for only four days (November 9-12).  And since November 10 was a Sunday, and November 11 was Armistice Day, this closure actually affected St. Louis's economy for less than two days.  No other American city or state in 1918 attempted even a brief lockdown of their economies.  In Los Angeles, some people argued for closing "all but essential businesses like grocery and drug stores," but after a prolonged debate over this, the City Council rejected the idea.

So there is no historical precedent in 1918, or at any other time in American history, or in world history for that matter, for what was done in March of this year--a government ordered shutdown of large sectors of economic and social life that has created the conditions for a global economic crisis, in which millions of people will be thrown into poverty, and millions will die from poor health and starvation.

As I have indicated in some previous posts, the one city with the lowest death rate in 1918 was not St. Louis but Grand Rapids, Michigan, which had two very limited closure orders lasting only 25 days. Retail stores, restaurants, and factories were never closed.  Churches were closed for only 18 days.

People in Grand Rapids believed that sick individuals should voluntarily isolate themselves at home, but there was no good reason to put healthy people out of work or to deprive people of their normal social and economic life, including public entertainment and church services.

In 2006, those planning a new strategy for responding to future pandemics--Mecher, Markel, and their colleagues--dismissed Grand Rapids as an "outlier" in the 1918 pandemic, and they looked instead to St. Louis as the model to be followed.  If there was ever another pandemic as deadly as the one in 1918, they argued, city and state governments should move quickly to order strict mitigation measures.  Sick or infected people should be quarantined. People should wash their hands frequently, wear masks, and keep some personal distance from others.  Travel should be restricted.  Social gatherings should be banned.  And all schools should be closed.

There were some critics of these proposals, who warned that the most severe mitigation measures could be "legally and ethically problematic," and they could be "devastating socially and economically" if they disrupted "the normal social functioning of the community."  Critics complained that there had been no moral cost-benefit analyses: "few analyses have been produced that weigh the hoped for efficacy of such measures against the potential impacts of large-scale or long-term implementation of these measures."

In answering these critics, the final official statement of the guidelines for mitigating a pandemic--first published in February of 2007--conceded that the likely health benefits of mitigation policies would have to be weighed against "the social, ethical, economic, and logistical costs" of such policies, so that the mitigation policies would not be justified if their likely costs exceeded their likely benefits.

The guidelines included a "Pandemic Severity Index" moving through five categories of severity.  The most severe mitigation measures would be justified only in response to a "Category 5" pandemic that would have a case fatality ratio and projected number of likely deaths comparable to the epidemic of 1918, which would mean projected deaths in the U.S. of over two million people.

For a "Category 5" pandemic, they recommended four levels of interventions:
"Home voluntary isolation of ill at home (adults and children) . . ."
"Voluntary quarantine of household members in homes with ill persons (adults and children) . . ."
"School child social distancing--dismissal of students from schools and school based activities, and closure of child care programs--reduce out-of-school social contacts and community mixing"
"Workplace/Community adult social distancing--decrease number of social contacts (e.g., encourage teleconferences, alternatives to face-to-face meetings)--increase distance between persons (e.g., reduce density in public transit, workplace)--modify, postpone, or cancel selected public gatherings to promote social distance (e.g., postpone indoor stadium events, theatre performances)--modify work place schedules and practices (e.g., telework, staggered shifts)" (p. 12).
It is said that the "hypothesis" that such interventions will really work to slow the transmission of a pandemic virus "remains unproven."  But it is also said that an "analysis of historical data for the use of various combinations of selected NPIs in U.S. cities during the 1918 pandemic demonstrates a significant association between early implementation of these measures by cities and reductions in peak death rate" (p. 29).  This report is silent, however, about Grand Rapids as the city with the lowest death rate, although it was not quick to impose such severe measures over long periods.

All of these recommended mitigation measures are familiar to us today as part of the guidelines for the COVID-19 pandemic.  But notice what is not here: nothing is said about shutting down "non-essential businesses."  Indeed, the term "non-essential" appears only once in the 61 pages of this document--in a reference to "non-essential social contacts" (p. 28).

A slightly revised version of this 2007 report was published by the CDC on April 21, 2017.  Like the earlier report, this one stressed the importance of "balancing public health benefits and social costs," which would require "estimating economic and social costs of NPIs and their secondary (unintended or unwanted) consequences," for the "minimization of social and economic costs during a pandemic" (p. 27).

In this 2017 report, there is no recommendation for shutting down "non-essential business" even in the most severe pandemic.  The term "non-essential" does not even appear in the document.

Beginning late in January of this year, some of the people responsible for the original 2007 pandemic planning report--including Carter Mecher and Richard Hatchett--started an exchange of emails about the possibility of a COVID-19 pandemic sweeping across the U.S. (see the link above).  By the middle of February, they agreed with one another that this was indeed the "Category 5" pandemic for which they were preparing in 2006, and therefore it was time to put their plans into action.

In justifying this conclusion, Mecher repeatedly points to the lessons from the success of St. Louis in the 1918 pandemic as compared with the failure of Philadelphia.  He says that in his public lectures, he regularly shows projected diagrams of the high death rates in Philadelphia and the lower death rates in St. Louis, and he then asks his audiences which city they would have preferred to be living in during the 1918 pandemic.  (See Mecher's emails for January 28, February 17, 28, March 4, 17.)  Apparently, he does not allow his audiences to see that Grand Rapids had the lowest death rates of any city.

Mecher insists that in response to this new pandemic, the benefits of NPIs far exceeds any costs.  He writes: "Back in 2007, there was modeling for estimating the economic impact of a pandemic (unmitigated with no NPIs) and a mitigated pandemic plus the costs of NPIs. . . . The bottom line is that when you add in the cost associated with lives lost in an unmitigated pandemic, additional healthcare costs due to greater numbers of those who are ill and hospitalized, economic costs due to lost productivity due to increased illness, the NPI costs pale in comparison" (March 10).

There is no indication here or anywhere in these emails that the calculation of "the NPI costs" includes the costs of a global depression caused by the massive shutdown of many sectors of social and economic life.  One might say, however, that those costs are not included because nothing is said in these emails about any need for such a massive shutdown.  After all, as we have seen, this was never part of the guidelines issued in 2007 and 2017.  Nowhere in these emails does anyone say anything about "non-essential business."

So who originated the distinction between "essential" social and economic activity that should be protected and "non-essential" activity that should be shut down during this pandemic?  And how did they show that the likely benefits of doing this would clearly exceed the likely costs?

On March 16, President Trump announced "The President's Coronavirus Guidelines for America."  The first line was "Listen to and follow the directions of your STATE AND LOCAL AUTHORITIES."  So Trump left it up to state and local governments to specify the legal orders implementing his guidelines.  The original plans in the 2007 federal statements anticipated that any legally coercive orders for responding to a pandemic would have to come from state governors and local governments.

The four levels of interventions recommended in the 2007 guidelines are included in Trump's guidelines--home voluntary isolation, voluntary quarantine of the sick, closing schools, and social distancing in workplaces and communities.

At two points, Trump's guidelines add something new to the 2007 guidelines.  First, in the last sentence, it is said that "in states with evidence of community transmission, bars, restaurants, food courts, gyms, and other indoor and outdoor venues where groups of people congregate should be closed."  Such legally mandatory closing of businesses was not part of the 2007 guidelines.

A second novel feature of Trump's guidelines is this sentence: "If you work in a critical infrastructure industry, as defined by the Department of Homeland Security, such as healthcare services and pharmaceutical and food supply, you have a special responsibility to maintain your normal work schedule."  On March 19, three days after Trump's announcement, the Cybersecurity and Infrastructure Security Agency of the Homeland Security Department published an "Advisory Memorandum on Identification of Essential Critical Infrastructure Workers During COVID-19 Response."  This seems to be the first use of the word "essential" applied to activity that must be protected in the pandemic response.  But the term "non-essential" is never used here.  And there is no recommendation in this document for closing "non-essential business."

This Homeland Security document on "essential critical infrastructure" was cited prominently in the executive orders issued by the governors of California, New York, and Pennsylvania on March 19.  Although Governor Newsom of California did not use the term "non-essential business," the idea was implied in his ordering "all individuals living in the State of California to stay home or at their place of residence except as needed to maintain continuity of operations of the federal critical infrastructure sectors," as defined by the Homeland Security document.  Governor Cuomo of New York was more emphatic: "all non-essential businesses statewide will be closed."  Governor Wolf of Pennsylvania ordered the closure of all "non-life-sustaining" businesses.  These orders by the three governors on March 19 were the first statements about closing "non-essential" or "non-life-sustaining" businesses.  

This kind of language became standard for most of the subsequent shutdown orders in the United States by the state governors.  Similar language has been used in the shutdown orders issued by governments around the world.

The social and economic costs of these shutdown orders are massive--including a worldwide loss of human life and liberty.  And yet, I have been unable to find any evidence that the people who proposed these orders demonstrated that the likely human costs of these orders would be less than their likely human benefits.

Moreover, I see no evidence that anyone in the United States involved in formulating these orders considered the possibility that they were unconstitutional.

So, again, I ask: When and how did public health experts demonstrate--through a rigorous cost-benefit analysis--that the COVID-19 pandemic justified a lockdown of our social and economic life leading to a global depression?  Or is it possible that our political leaders have rushed into these lockdown orders without the support of any such moral reasoning?


Alexander Gieg said...

The methodology for this cost-benefit analysis hasn't been developed until very recently. It's now available, the authors of the paper having developed a new metric called "Wellbeing-Years" so as to be able to compare very disparate factors, from death rates to economic impacts:

Clark A, De Neve J-E, Fancourt D, Hey N, Krekel C, Layard R, O'Donnell G (2020). "When to release the lockdown: A wellbeing framework for analysing costs and benefits." Occasional paper 49, CEP Occasional Papers 2020-04, Wellbeing Programme at the Centre for Economic Performance, LSE.

Larry Arnhart said...

Thanks for this. I will want to read this paper carefully. After glancing over it, it seems that they have not taken into account that the UK lockdown has a higher pandemic death rate than Sweden with no lockdown.

Alexander Gieg said...

You're welcome. I found the link in this very detailed summary covering everything known so far about the illness, so you might be interested in it too. It's a very long read, and well worth following its eventual updates: Peter's COVID Consolidated Brief - 29 Apr.

About the paper, it's purpose, as I understood, is more one of presenting the method than providing definite results for the UK case. The researchers offer a sample calculation based on known facts and some assumptions, but they say the method itself should be applied to each country, and their subdivisions, as accurately as possible to obtain reliable results. As such, a detailed calculation Sweden would be needed for Sweden before any comparison between both could be actually provided.

Naimisha said...

Dear Dr. Arnhart -fyi, I link to and draw upon this fine post in my piece "The Mystery of the Lockdowns" -