“It became necessary to destroy the town to save it”: The case against social distancing
March 21, 2020 |
Below I am posting a piece written by a fellow Ontario emergency physician on the front lines of the COVID-19 battle. It’s an interesting counterpoint to the numerous circulating emails from healthcare professionals pleading for some form of panic. Panic helps no one, and our success in fighting this pandemic will depend on honesty – with ourselves, and with the data.
He believes that we have every right to be angry with the current prescription – one that shifts responsibility from our overburdened healthcare system to ordinary citizens. But more importantly, that our current strategy will not save lives. True compassion means acknowledging what works and what does not, that some will get sick, and that we will certainly need the resources to care for them in the trenches.
I called my 76-year-old mother this morning after my night shift in the Emergency Department. She told me that she tried to go to church yesterday, but there were only two people left in the pews and daily mass had been cancelled. Social distancing, they told her. She got down on her knees and cried.
On March 14, everything changed. The United States closed its borders and declared a national emergency. The stock market crashed by more than thirty percent. All schools, sports, restaurants, bars, and churches in Ontario were shuttered, and an advisory was announced against all international travel. In an instant, the fiat of public health authorities stripped away most of the meagre joys of human existence and replaced them with the vague promise that they might be restored at an indefinite time in an uncertain future.
Through their actions, public health decision makers have made clear their belief that social isolation is a largely abstract concern when fighting a threat as existential as COVID-19. But like most things in life, our approach to COVID-19 involves an impossible balance of risks and values – on one hand, the risk of an elderly woman going to church in the midst of a pandemic, balanced against the competing value of “How long will it be before we can hug our loved ones again?”
I intend to argue that with indefinite lockdowns and extreme social distancing, we have lost sight of this balance. Unlike the successful Korean approach to suppression, we have implemented unrealistic and unsustainable measures in an atmosphere of amplified panic – and in the absence of good evidence.
1. What is the mortality (death rate) of COVID-19?
I don’t know.
But then again, neither does anybody else. What we do know is that the number of cases are certainly underreported, and that actual mortality is likely low, but higher than the seasonal flu. That doesn’t mean we can’t make an educated guess.
Let’s examine four countries and their mortality on March 18 (Source: WHO Situation Report 58):
Korea 8320 (81 deaths; 1.0% mortality)
Afghanistan 22 (0 deaths; 0% mortality)
Italy 31506 (2503 deaths; 7.9% mortality)
China 81116 (3231 deaths; 4.0% mortality)
The numbers are spectacularly different. Then which one is true? Well, Korea. Sort of.
We don’t believe China, because we never ever believe China. On its way to 80 thousand cases, China locked itself down in a way that only an authoritarian dictatorship can do – and in so doing, suddenly stopped reporting cases of the COVID-19 virus. I am told that four months after the outbreak, life has mostly returned to normal. If this is because most of China has been infected with the virus (and is now immune), then 3231 deaths spread among several hundred million people doesn’t sound quite so terrifying. But if most Chinese have not yet been infected, then we should expect another dramatic second wave of infections among those who are not yet immune. We don’t know yet. But more likely, we simply should be extremely skeptical of China’s numbers.
Similarly, we can’t believe the Afghan numbers. This time, it is because they are too small, and because the ability of their health authorities to accurately keep records is questionable.
The appalling Italian numbers, on the other hand, are a gross overestimate of mortality. That’s because of the tendency to test the sickest people with the worst prognosis. In general, what we usually see is that the more people who are tested in the community, the lower the apparent mortality becomes. For example, Korea tests 700 times more people per capita than the United States, and their numbers are much more appealing.
But the Korean numbers are an underestimate of mortality. Many of the confirmed cases are still open cases – in other words, they are patients who may eventually die from their illness. For COVID-19, it appears that this happens around day ten of illness. So, to calculate the true mortality, we must compare the deaths today with the confirmed cases ten days ago. For Korea, that means 7134 cases on March 8 and 81 deaths on March 18. That’s a mortality of about 1.1% for confirmed cases.
And even that is not the whole story.
The number of confirmed cases is an irrelevant number that bears little to no relationship with the actual number of cases. Otherwise, the best way for a country to entirely eradicate its COVID-19 epidemic would be to run no tests at all. The truth is that for every confirmed case, we have no idea how many asymptomatic and minimally symptomatic cases there are walking around our neighborhoods. Most estimates are around 5 to 10 times as many (Source: Science, Li et al. 2020). There is some suspicion that children may even act as asymptomatic disease vectors. Also, consider that our diagnostic test for the COVID-19 virus is only around 75-85% accurate (Source: Young et al, 2019).
What this means is that the actual mortality for COVID-19 could be as low as 0.1-0.2%. We don’t know for sure. But those aren’t numbers that should inspire panic in the average individual.
It also means that the only meaningful numbers are admitted patients, critical care patients, ventilated patients (those on a breathing machine), and deaths. Those numbers are important because they give some sense of the burden of the disease, and because they cannot be fudged. But other than the death count, these key indicators are surprisingly difficult (although not impossible) to find in any published form:
When the WHO and public health authorities sow panic by telling you that mortality for COVID-19 is 3.4%, that should make you angry.
2. Should we test more, or test smarter?
Smarter, always smarter. Despite the widespread complaints in the North American media that not nearly enough people are being tested.
Bo, a village in Northern Italy, tested all 3300 inhabitants and found that 3 percent of its population was infected, of whom half had no symptoms at all. Even after isolation and retesting, asymptomatic carriers remained, any of whom could have begun a second wave of COVID-19 infections.
Community transmission is already established in most of the world, including Canada. That makes the testing and retesting of 38 million people unfeasible at best, and harmful at worst. Harmful, because it means that anxious but otherwise medically well patients might flood our emergency departments and clinics. In doing so, they expose themselves and others to infection, and consume valuable human and physical resources as healthcare workers gown, glove, and mask to assess them. This is not the approach that Korea takes, of course – they’ve been running up to 20 000 drive-through tests per day on anyone who requests one – although it’s not entirely clear to me how this is any more efficient than forcefully telling anyone with symptoms to stay home.
The reasons to test anyone for anything are short indeed. First, we test if there is a specific treatment available, which for mild cases of COVID-19, there is not. Second, we test to protect healthcare workers and vulnerable people. And third, and perhaps most importantly, we test for disease surveillance.
As I wrote above, numbers of confirmed cases are a meaningless statistic. Instead, we should be following the prevalence of the COVID-19 virus in admitted respiratory patients. This is the same approach that we already take with flu surveillance (Source: Government of Canada FluWatch). It also allows us to map out the proportion of patients with the pathogen – one that indicates whether progressively more, or fewer ill patients are being admitted to hospital with the COVID-19 virus.
Despite that, it presently takes a full four days for Ontario to turn around a COVID-19 diagnostic test for an admitted patient, even one who is critically ill. That, too, should make you angry.
3. But will it overwhelm our healthcare system?
Well, maybe. And that should make you absolutely furious.
There are around 400 000 flu cases each year in Canada, leading to anywhere from 300 to 4900 annual deaths. In a bad flu year, that number could spike as high as 8000 deaths. For simplicity, if we speculate 30 million people in Canada, a 30% COVID-19 infection rate, and just 0.2% mortality, that’s already 20 thousand deaths. That could make COVID-19 the mother of all bad flu years.
If even 1% of those COVID-19 patients are sick enough to need critical care, that would require about 100 000 ICU beds. Each would be needed for 5 days or more. Consider that we have 5000 ventilators in Canada. Most are already in use, but we will pretend they were not. Even if they were not in use, it would take three months to adequately service this number of critically ill patients. Play with the numbers if you like – mine are more optimistic than many other figures you will read.
Why don’t we have enough ventilators? It’s a good question. Even prior to COVID-19, it was already rare to call the ICU with an admission without learning that my patient would be taking the last remaining bed. Sometimes, when there are no beds at all, a patient will remain in the Emergency Department for hours, or even days.
The fact is that we have been through all of this before, during SARS in 2003 and the H1N1 swine flu pandemic in 2009-2010. In fact, a coronavirus has long been predicted to be the cause of the next pandemic. The fact is that we have been operating a straining healthcare system at 110% of its capacity for years – it’s no wonder that it has no capacity to surge in times of crisis.
None of this has ever been a secret. The fact that we were not remotely prepared should make you angry.
4. So that means we should ‘flatten the curve’ with social distancing, right?
Now, we arrive at the greatest of the lies we have been peddled. Having failed us at every step along the way, the healthcare system demands instead that the average citizen attempt to defeat the virus by sheer force of will.
Social distancing seems to have appeared in the lexicon overnight. And while social distancing is by no means a new concept in public health, the scope of the lockdown that is being attempted today is unprecedented in modern times. It is a social experiment with consequences that are certain to be catastrophic.
Flattening the curve theoretically looks like this in most media sources. Note the conspicuous absence of any units of time on the x-axis.
Let’s find a flatten the curve graph that actually has numbers, these ones for the United States and UK. The numbers on the y-axis are truly staggering, but then again, so are those on the x-axis:
If we refer back to my previous estimates, the idea is to spread our 100 000 ICU patients over an extended period of time, in order to avoid a surge in infections that overwhelms the healthcare system. On paper, such a strategy has the potential to decrease the overall size of the epidemic. Most people I have encountered believe that period of time to be around two weeks. They are dead wrong.
As I have argued above, even if every ventilator in Canada were idle and waiting for a COVID-19 patient, the critically ill patient load would need to be spread over a period of three months. But given that most ICUs already operate at or above capacity, a more likely optimistic scenario would be six months to a year. Against a six-month horizon, a two-week token effort would be trivial. To pretend that two weeks of social distancing is anything more than a symbolic gesture is disingenuous. Without guns, telling people that they cannot work, gather together, nor visit their loved ones is simply unsustainable over a period of months. To do so indefinitely will leave the world unrecognizable.
That is, even if we take the effectiveness of a lockdown against COVID-19 as a given. I am not a public health physician, and therefore, I will entirely leave this section of discussion in the hands of those with greater expertise than my own.
If social distancing is to be used to control disease, it is very important to get the parameters right. If the infectiousness of the disease is underestimated even slightly, the actual outcome of using the control can be substantially worse than anticipated. The control becomes less effective for more infectious diseases, resulting in a worse outcome than doing nothing. If a disease is very highly infectious, social distancing may have no effect, or may require an unfeasibly high degree of caution in order to be effective. In these cases, doing nothing will be a more cost-effective strategy than using social distancing, because the worst-case outcome arises if the control is applied, but the level of caution used is too weak.
In other words, an ill-considered social distancing intervention can be worse than having done nothing at all. The effectiveness of such distancing depends on two mathematical variables, neither of which are known to us. First, the risk attitude of people in the community (i.e. how frequently they break the draconian rules), and second, on the infectiousness of the disease. Not only do we know neither of these variables with any degree of accuracy, but what little we know points to both being hopelessly stacked against us.
First, all existing evidence points toward COVID-19 being an extremely infectious disease. It’s likely present in asymptomatic carriers, children, stable fomites, droplets, and may even be aerosolized (Source: Doremalen et al, 2020). Viruses are brilliant – they know that keeping people mostly healthy is the best way to spread infection. It’s a virtual certainty that I am exposed to the virus every single day that I go to the hospital.
Second, since asymptomatic people are unlikely to avoid their loved ones, parents, friends, and spouses – and will certainly not do so for six months – this means that most likely result of all this sacrifice is that our attempts at extreme social distancing simply will not work. At all. The stricter the lockdown, the less time it can last before it becomes the chemotherapy that kills the patient. If we’re simply buying time for the cavalry to arrive – ventilators, beds, human resources, protective gear – we need to know now.
Because the corollary is that the moment that widespread poverty, loneliness, and boredom spell the end of this experiment, we will experience the exact same immediate spike in infections in the unexposed that we had hoped to kick down the road. In the 1918 Spanish flu pandemic, suppression was transiently effective, but if insufficient herd immunity had developed, infections often rebounded as soon as controls were lifted (Source: Bootsma, 2007).
Ours is an experiment with no control set, undefined variables, multiple confounders, and no defined outcome measures for its effectiveness or failure – and therefore minimal accountability. The psychology of this amounts to classic game theory.
With no malice intended, policymakers have set themselves up for success. First, they sow panic by predicting deaths on par with the first World War. If the pandemic unfolds more favorably than the apocalyptic predictions, they can claim victory. And even if we are buried by an epidemic of catastrophic proportions, they can claim it would have been that much worse without intervention, that the public lacked sufficient resolve, or that authoritarian controls were not tight enough.
These claims are neither testable nor falsifiable. They are articles of faith.
5. Okay, but it’s worth a try, if we can just save one life, right?
No, because the collateral damage of doing so for six months is unfathomable. It’s a pound of prevention for an ounce of cure. If it weren’t enough that locking down the population may be ineffective, it is also simultaneously inefficient, impractical, and inhumane.
First, there’s the economic devastation to those who lose their businesses and their livelihoods. I am no economist, but I am told that the American stock market has wiped out $11.5 trillion in value since its peak in 2016. That’s enough to buy 230 million ventilators, or one $50000 ventilator for nearly every man, woman, and child in America. That is to say, the enormous resources being poured into this hopeless battle for containment could and should be repurposed to damage mitigation.
The healthcare costs of extreme social distancing are not abstract, nor are they insignificant. Social distancing further isolates the marginalized and the elderly, particularly those who live alone, causing loneliness, depression, and increasing mortality (Source: Holt-Lunstad). It limits physical activity and worsens anxiety, mental illness, suicidality, alcoholism, and substance abuse. It’s only the early days, and I’m told by police officers that domestic violence has already spiked. While Korea quite reasonably shut schools and banned mass gatherings, in North America and Europe each and every social activity that provides value and meaning to people’s lives has been stripped away. None of these harms are hypothetical, nor are they any less valid than the laser-like focus of public health on pandemic containment. Make no mistake – to do so for 6 months is impossible, and it is of no use asking the impossible.
No public health intervention with such disruptive effects on society has been previously attempted for such a long duration of time. How populations and societies will respond remains unclear.
The world of COVID-19 is a petri dish for human vulnerability – fear of illness, fear of death, and fear of the unknown. And so far, we humans have responded with a chorus of groupthink, deference to authority, and increasingly, shaming and tests of the purity of each other’s religious commitment to social distancing. Instead, we should be responding with anger – anger that China is somehow hailed as a role model for disease containment. Anger that France has moved to require its citizens to complete a permission form each time they leave their home. Without a hint of hyperbole, we risk complete societal disintegration, martial law, and the loss of our essential civil liberties.
I could end by suggesting that the greater the impositions on personal freedom, the greater the skepticism with which free people should react – and that would be true. I could tell you that life itself has never been safe, and never will be safe. But it’s equally true that I don’t want my mother to cry over the phone, and I want her to be free to go to church again.
A contrarian prescription
Don’t sow panic. Instead, reassure the public that while COVID-19 is a genuine public health threat, their individual risk is low.
Don’t close restaurants, cafes, and other small businesses. Instead, shut schools and limit mass gatherings, reasonable limits on behavior that have the benefit of being sustainable over a period of many months.
Don’t close borders. We already have community spread in Canada, and so the contribution of imported cases is now largely irrelevant.
Don’t broadly test in the community. Instead, make it abundantly clear that all cold or flu symptoms should be assumed to be COVID-19. Since our tests are imperfect, those with respiratory symptoms should be told to self-isolate and supporting sick workers in staying home should be made an urgent public health priority.
Don’t set frontline workers up for failure. Instead, ensure that Canada has an adequate stockpile of ventilators and protective equipment for inevitable pandemics like this one.
A concerned Ontario Emergency Physician
Sign up for free to flag trips of interest and email organizations directly through our directory.