N/B: This is a podcast script for the first of our episodes about the COVID-19 epidemic. It first came out on 17th March 2020. For that reason, much of the scientific and medical information, though accurate at the time below, has been superseded by new information. Please read this as it is — a historical document composed in the midst of a breaking news story, which has not been edited.
We have done episodes subsequently on the COVID-19 pandemic, but I won’t be publishing the scripts unless there is demand for them — in which case you can contact us via the web at physicspodcast.com. All of our episodes on the COVID-19 pandemic can be accessed here:
Hello, everyone. Did you miss me?
For the last few weeks and months — while I’ve been struggling with juggling work and some ongoing health issues — I’ve had this internal debate going on in my mind about the podcast. I really wanted to bring it back, as soon as I had the time and energy to do so. At the same time, all of the energy I’d usually use for research outside of my ordinary work was entirely focused on the threat from what was called the Wuhan coronavirus, but is now very much the world’s coronavirus.
If you’re listening to this, you probably know all about the basic facts. We are now in the midst of a global pandemic, the potential scale of which we haven’t seen for a hundred years since the Spanish Flu in 1918–1919. If your country hasn’t yet implemented lockdown measures, they are probably on the way. As I write this in London, the official advice is to avoid all non-necessary social contact, and in a few weeks, the advice may well be even more draconian.
So I was stuck in the midst of a debate. Do I write an episode about the coronavirus pandemic? On one level — I’m a physicist, not an epidemiologist, and so it’s not my area to write a special episode on this topic as it was when, say, the Neutron Star merger happened. I’m totally aware that there is massive saturation, everywhere, about this pandemic right now. A lot of what I have to say will not be news to you, and will just add one more “take” to an ocean of “takes” that you can find across the planet at the moment. At the same time, the information we have is rapidly being updated, all the time — and so this podcast, even if I get it finished and recorded today, will probably be out of date a few days or weeks after you hear it. This is probably already far too late, and I should have written and recorded it weeks ago when I first started researching to get the information out there, but that doesn’t matter now: my fault for being slow. Some of the information may be wrong, and I’ll attempt to point out where I am guessing. Not even the experts can truly, honestly give you more than a very well-educated guess about how this is all going to pan out, because a pandemic on this scale with this kind of virus and this many unknowns is quite unprecedented in their lifetimes too, and in modern society. So not everything will be correct. And the psychological burden of dealing with this thing for what I’m afraid will be many many months to come is going to be severe for all of us. It may well be that the last thing anyone wants is yet another discussion of the coronavirus.
These are all excellent arguments for me keeping my trap shut and instead continuing to work on the other scripts that I’ve been squeezing into my spare time lately. But on the other side — on the other, overwhelming, compelling side — is the fact that, since about the 12th of February when the first case arrived in my country (the UK) and it became very clear to me that this was unlikely to be contained within Wuhan and within China, where I’d been seeing terrible messages from for weeks since the lockdown– it’s been pretty much all I’ve been thinking about, reading about, conjuring up worst-case scenarios about.
So, if the compelling arguments not to make a coronavirus episode sounded good to you, switch off. If you are going to find it upsetting to listen to me discuss this as a supposed leisure activity, switch off. Call a loved one instead. I will try my hardest to be back soon with what I promise will be non-pandemic related content about, I don’t know, stars or artificial intelligence, or anything aside from the pandemic. If for some reason you’re a sucker for punishment on this stuff, indulge me as I try to write about it. This is a therapeutic exercise for me as much as anything else, just to try and clear some space in my brain for — well, almost anything else.
I should also point out that a lot of what drove my concern around this virus was all the research I did way back in 2017 for the TEOTWAWKI series — The End of the World As We Know It — on pandemics. I listened back to that, and while bits of it sound hopelessly naive and scientifically muddled, other parts of it sound more than a little prophetic. The whole message that the relentless research on “the end of the world as we know it” really drove home to me is just how much we all seem to accept that our way of life, our way of understanding the world and behaving in it, the structures that form our society… are taken for granted as immutable, unchangeable, when in actual fact there are plenty of things out there that might change them. Both on a personal and a global, societal level, there are rare events that can happen which will just wreck all of the assumptions you made about the future wholesale and cause you to completely re-organise your priorities. For the last few weeks as the news has continued to accelerate I’ve been conjuring up this kind of scenario in my head… and hoping that it wouldn’t actually come to pass, in the end…. But that is what is happening now.
I should also say that by no means is this “the end of the world” — but the world won’t be “as we know it” for quite some time. One of the curious things about this pandemic is that, statistically speaking, the risk to any given individual is actually probably quite low, maybe lower than you imagine it being, for reasons that we’ll get into further a bit later on. Unfortunately, when you multiply a low risk to any given individual by a hell of a lot of individuals, the risk to society at large becomes substantial. 1% multiplied by 70 million people is 700,000 people. And again, if that’s the kind of sum you don’t want to hear about, switch off now.
So let’s begin with what we know.
In December 2019, the first cluster of pneumonia patients of an unknown cause appeared in a hospital in Wuhan, China. Wuhan is a city of about 11 million people — so about the size of London — and it’s just a testament to how huge China is that this is not a city that a lot of people had heard of prior to this outbreak.
Epidemiological tracing has worked out that the first person to apparently have symptoms had those symptoms on the 1st of December, although there is a lot of anecdotal evidence that suggests that the virus was circulating in people substantially before that. So this is one of the first, the key things to understand about this virus, I think — there’s a delay involved. Once you get infected, it can take up to fourteen days for any symptoms to show up. Once you get ill, the disease progresses like a mild cold or flu in most people for the first week or so, and in a small fraction of people who get it, then continually gets worse until they have this viral pneumonia — that’s inflammation of the lungs that is caused by a virus.
So the most obvious sign that anyone has the disease is when a whole bunch of patients with viral pneumonia show up at the hospital, and they test negative for the flu or anything else that could cause this kind of illness. But you can immediately see why this is a problem. If, say, 10% of people who get this develop pneumonia and symptoms bad enough to go to the hospital — and we’ll come back to this later — then let’s say you notice 10 people with viral pneumonia showing up at your hospital. That means that three or four weeks ago, you had 100 people in the community with the virus — most of whom will have been spreading it for a while. By the time someone unfortunately dies of this disease, statistically speaking, as the death rate with good medical care is around 1%, you probably had 100 people with the illness in your community as long as a month ago. So by the time you’re aware of the outbreak, it’s already bigger than you think it is. In fact, in the pandemics episodes in 2017, we theorized that the worst kind of virus to emerge would be one that was initially asymptomatic or showed small, hardly noticeable symptoms for a long time — which would allow people to travel and spread it internationally before anyone realised what was happening — and, unfortunately, that appears to be exactly what has happened here, which explains why the situation is currently so severe.
Some terminology, then; this virus is called SARS-COV-2, for reasons we’ll explain, and the disease that it causes is coronavirus disease 2019 or COVID-19. So the virus itself has a different name to the disease that it causes; like most people, I’ll just carry on saying coronavirus.
Initial reporting suggested that the cluster of cases of this new virus could mostly be related back to a seafood market in Wuhan itself. This would make sense as the origin of a new kind of virus, because, as we discussed in the pandemic episode, way back in 2017, new kinds of worrying viruses often arise from animal transfer into humans. In these markets, lots of animals and lots of humans are living in very close proximity, so it’s easy to imagine how a virus that began in bats, which are mammals, might mutate and jump into other mammals, like pangolins, and then mutate once again and make the leap into humans. This is the current working theory for what might have happened here. Bats are reservoirs — they have potentially hundreds of different strains of viruses in their system at any one time — so while the probability that such a mutation happens is not all that high in isolation, with enough time and enough mutations, eventually, this kind of thing does happen.
And indeed there were two previous outbreaks — one of Sudden Acute Respiratory Syndrome, or SARS — and another of MERS, Middle East Respiratory Syndrome — which have also been traced to coronaviruses that are suspected to have crossed over from animals into people at one point or another. These outbreaks make for good case studies as to why some viruses are more potentially dangerous to cause massive pandemics than others. Both of those viruses have much higher fatality rates than the new coronavirus. SARS ultimately killed around 11% of those infected. MERS was around 34%.
Paradoxically, the high fatality rate is a big part of the reason why these viruses have been easier to contain so far. It is actually, paradoxically, the fact that the new coronavirus has a low fatality rate and appears to be mild for most people who get it that is allowing it to spread so quickly. The traditional approach to containing such an infection is one that you’ve probably seen your own country attempt to do in recent times; test sick people, work out who they came into contact with, trace each and every one of those contacts and quarantine them, providing medical treatment where required. For small outbreaks, you can actually hope to cut off nearly every connecting strand to prevent the virus from getting out and infecting anyone else. A few weeks ago, it became clear in most European countries that “community spread” had started, i.e. cases that were not linked to any known other cases, which is a sign that this attempt to contain the virus was failing. It has now failed in most countries in the world, although it was always going to be an extremely difficult task due to the basic properties of the virus, and let me try to explain why.
People infected with SARS or MERS were usually severely ill — very easy to identify. Also, if you’re severely ill, you’re unlikely to be going out and infecting other people in the community and much more likely to be isolated at home, where the contacts you have are easier to trace, or in the hospital. Not that outbreaks can’t occur in hospitals — the Toronto SARS outbreak, which ultimately killed 43 and resulted in 251 cases — was traced entirely back to a single patient who went to hospital after flying back from Hong Kong with the illness. A big part of the reason why was that few people were aware of the condition and they didn’t have the proper protective equipment while treating the patient; in the Prince of Wales hospital in Hong Kong in the early days of the outbreak, up to 99 people were infected by a single patient while treating the person. There are reports that people with SARS were not actually really contagious until a few days after they started showing symptoms; which means that, as long as everyone is on high alert, you can hope to track down and identify each individual case of SARS and limit the spread to healthcare facilities. MERS, where the disease’s severity is even worse and it appears to struggle to transmit person-to-person as easily, is easier to contain than the new coronavirus for similar reasons.
In the pandemics episode, we pointed out that viruses like Bird Flu — where the fatality rate is high and the symptoms often severe — are thankfully not well adapted to become pandemics for this very reason. As a virus, it’s best not to kill or incapacitate your host; viruses need hosts to survive. If they can walk around essentially perfectly happily infecting others for an extended period of time, as is the case with common colds, the virus won’t die out so easily. New viruses that transfer from animals — we call them zoonotic in origin — are often not particularly attuned to humans, though, which is why initially they can kill an awful lot of people, and sometimes will eventually be replaced by a milder strain that can spread more easily, as it’s actually evolutionarily advantageous for the virus not to kill people, but to continue to reproduce for as long as possible.
With the new coronavirus, though, we have no such luck; lots of cases are mild, and the disease spreads rapidly because, as it’s a new illness which no-one in the world was immune to (and even now, very few people in the world are immune to this virus.)
This means that we need to look at some of the characteristics of this virus to understand what this pandemic might potentially result in. One of the key measures that people use to characterise pandemics is the so-called R0 value, the basic reproduction number. This is, if you’re infected, the average number of people that you will go on to infect. It’s worth saying that this depends on an awful lot of things. For example, it depends on the symptoms that the virus has, how good it is at spreading itself around, how long people are sick for, and also how many people they interact with.
Initial estimates for this virus suggest that the R0 is somewhere around 2.4, perhaps between 2 and 3 to give decent estimates. One obvious point to make here is that as long as R0 is more than 1 — i.e. each person infects more than one other person on average — you’re going to see a growth in cases. The aim of containment and control measures is to reduce R0 to less than 1. When we talk about R0 here, it will be assuming that there are no control measures at all.
We can compare this to other diseases. Seasonal flu has an R0 of 1.4 — it’s less contagious, perhaps because people have some immunity to seasonal flu due to the vaccines that are regularly produced, encountering similar strains before, or perhaps due to the nature of the symptoms for a bout of seasonal flu. But there are other diseases that are far more contagious — measles, for example, has an R0 of between 12 and 18, and smallpox had an R0 between 5 and 7, so you can see that there are some truly terrifying pandemics out there in history that have been even more infectious than this disease.
We can learn something else interesting from R0, and this is that it actually gives you a rough estimate for what total percentage of the population might end up being infected with a disease if nothing is done. This very rough estimate is quite easy to understand. Let’s assume that once people are infected, they are immune. If your disease has an R0 of 2, then it means that you will infect two other people on average. Imagine that the whole population is mingling about, getting sick, and recovering, until half of everyone is now immune to this disease. Then, of those two people you go on to infect, one will be immune already, and won’t get infected. So the actual R0 is now 1 — which means that you’re infecting just 1 person on average. At this point, the number of new cases will start to decline.
This is something a little like what we might call herd immunity; if you get infected, but most of the people you meet are immune, there will be no outbreak. For a very simple virus model, we can expect 1–1/R0 of the people could get sick in total, and that will give you, with R0 2.4, about 58% of people getting sick in total before the disease will stop spreading.
Now clearly you see a couple of things. The first is that this is a really rough estimate. If people don’t widely circulate, or say infected people are kept together and away from immune people as much as possible, then the R0 is going to go down. It’s assuming no control measures and everyone mixing together. The second is that this is an equilibrium estimate. Physicists love these! Once you can take time out of the equation, by just looking for the next time your system will stop evolving and be in a steady state situation, you don’t need to worry about that pesky time dependence any more and can just consider what things will look like in the long term. In equilibrium, once the disease has run its course, in this toy model, 60% of people could be infected. And finally, you can see just how crucial uncertainty about the true R0 value is to how much damage you think this disease is going to do. So for example there is a paper which you can read online called: Estimation of the reproductive number of novel coronavirus (COVID-19) and the probable outbreak size on the Diamond Princess cruise ship: A data-driven analysis.
They estimate there that the R0 is probably between 2.06 and 2.52. Initial estimates with less data assumed it was between 2 and 3. But the difference between 2 and 3 is the difference between 50% and 67% of your population eventually getting sick; which is millions of people in your country or mine. For the seasonal flu epidemics you might naively expect 28% of the population to get sick if there was no vaccine at all for a new flu.
You can also, incidentally, see from this toy model why vaccines for much faster-spreading illnesses like measles are so absolutely crucial. If measles has an R0 of 12, then to be safe, you need to have 92% of the population vaccinated — otherwise, if anyone gets measles, they will probably spread it to more than one other person and you’ll have an outbreak that will grow in size, although obviously not as badly and not as quickly as it would if no one was vaccinated at all.
There’s one other important thing to talk about for this toy model, and it brings in another toy model that is the favourite of physicists and tech people everywhere. I noticed that a lot of silicon valley companies were the first to close down their events, and I actually think this is why: they think exponentially.
Exponential growth in mathematics, you can simply think about as doubling in a given period of time. For this disease, we suspect that doubling could occur every 4–7 days, depending on how much people are interacting and so on. And this means, in practice, if you’re not used to thinking about these things, that stuff will very quickly appear to get out of hand. So let’s put some numbers to that. Assume a doubling time of 4 days and we start with a single sick person; these are just estimates, reality is probably a bit quicker than this. After 8 days, we’d have 4 sick people. After 16 days, 16. Barely noticeable at this stage. After a month or so, we’d have 256, at which point you might start to have a few cases of pneumonia show up at your hospital. Another week, 1,000. Another week, you’d have 4,000 sick people, and you’d start to notice a few people dying at your hospital of coronavirus. That would be around six weeks from the start of the epidemic. After another two weeks, two months or so in total, you’d have 65,000 sick people. It’s worth noting that according to the UK Chief Medical Officer, this is approximately where they think we might be now as I write this on 17th March. This is pretty coherent with the epidemic starting with a single undetected patient in mid-January, which makes sense, as people were likely travelling in from Wuhan at that point. You can infer the numbers just as easily for your own country. They will obviously vary from this toy model — maybe the first arrivals from Wuhan kept to themselves, or maybe they hopped straight on public transport; maybe efforts to track down early cases were successful at slowing the spread, or maybe they weren’t — a lot of luck and randomness comes into this, but it should on average be approximately correct. But remember that because pneumonia takes 2–3 weeks to develop, right now, the hospitals are only seeing the results of what we had 2–3 weeks ago. Essentially, the pressure on hospitals now is from the 4,000 people who were sick 2 weeks ago, not the 64,000 who are probably sick now.
Naturally, you can then carry this on to its grim conclusion. You would have, in another two weeks or so, perhaps 500,000 sick people; maybe a million a week later; and so on, and so forth. At this point, things will get rather grim for your healthcare system.
One thing I haven’t mentioned so far is testing. Testing is clearly crucial to any attempt to squash the epidemic in the early stages. But testing is inevitably limited, and not just by things like the number of test kits you can produce. For a test to be properly conducted, someone needs to come to you in full protective equipment and swab your nose and mouth. With limited nurses and medical staff who can do this at any one time, the bottleneck for conducting tests is going to be in trained personnel rather than necessarily anything to do with the number of tests that have been manufactured. Today, the UK managed to test around 4,000 people, which is pretty good going, but it’s obvious that the testing now is just going to give us an idea of which regions and individuals are being hit, rather than necessarily doing anything to stop the spread of the virus. If the rough exponential model is accurate — and it appears to be close to the estimate of 55,000 that the Chief Medical Officer gave today to parliament — then obviously the 2000 people who have tested positive in the UK are a drop in the ocean, and we can no longer rely on the classic SARS method of testing people, isolating them, tracking down their contacts, and so forth — although this is all valuable work to slow down the spread of the disease if it can possibly be done, it probably won’t now contain the epidemic.
Given that different countries are testing by different amounts, one very rough — if morbid — way you can try to infer how well each country is doing in its outbreak is with the death rate. If you assume that the death rate is somewhere between 1 and 5% depending on healthcare — of which more later — you can then divide the current number of deaths by that, and account for the time lag, to get you an estimate for how many people are sick now. In the UK, as of today, 55 people have sadly died. That means that three weeks ago when those people started to get sick, we probably had around 5500 cases. So with a few doublings, you can see that we probably now might have between 44,000 and 88,000 cases, which again is what the Chief Medical Officer briefed to Parliament earlier today. So these aren’t magical numbers that are difficult to figure out; it’s hard to be precise, but in terms of the order of magnitude, the power of ten, we can easily infer these numbers on our own.
So in terms of the basic principles for me understanding this epidemic, there are these. We have the time-lag principle — because people take time to show symptoms, and even longer to get sick enough to go to the hospital, whatever stage you think you’re at, you’re likely a few weeks further along in terms of people who have actually been infected. We have the idea of exponential growth that is potentially only limited by the basic reproduction number R0 and the fraction of people that are going to get this disease. Now I want to talk about another key principle in understanding this epidemic that I’ve realised through all this obsessive reading about it, and this is what I’ll call confirmed case bias.
So the point here is essentially a simple one. Most people who are being tested for COVID-19 will be sick. Many of them will be very sick — sick enough to need to go to the hospital. Most cases of COVID-19 that are observed by medical professionals will be the most serious cases, because they will be the ones going to the hospital. If you have a mild cold or flu-like symptoms in Wuhan at the peak of the epidemic, or Italy now, or soon enough anywhere else in the world, you are not going to be able to be admitted to the hospital; there will be too many other people who need it.
This means that frontline healthcare workers just see, for the most part, the sickest tip of the iceberg. In the limit, imagine an extreme case where a disease was totally undetectable for 99.9% of people and killed 0.01% of people almost instantly. [This might actually be true of some diseases but I feel like we’re all such hypochondriacs right now I won’t go into that any more.] All the cases detected by hospitals would be the ones who died straight away, and we’d assume that the disease had a huge fatality rate. The reality with COVID-19 is somewhere between that and what we might hope would happen.
So let’s talk quickly about what the hospitals are seeing. For this, I will refer to the Chinese CDC study which focused on the outbreak in Wuhan. Due to the extreme containment measures by Chinese authorities — and I mean extreme; people have not been allowed to leave their houses except for groceries for months, people who violate quarantine arrested, people required to wear protective equipment in public at all times, etc. — the initial outbreak there is essentially over, so it’s actually a decent case study for what a completed outbreak might look like.
The hospitals in China have observed that, of the patients admitted, typical symptoms looked like this:
Shortness of breath
Nausea or vomiting
I think this is really important information to know. I had a cold a couple of weeks ago, the main symptom was a hilariously runny nose, so I assumed that it wasn’t coronavirus because it would have been unlikely to appear without a fever (and also the fact that my estimate for the maximum number of cases in the UK at that point was very small.)
The other thing that the doctors observed, and this is really crucial, is that approximately 80% of cases were mild; they just looked like ordinary flu, so not nice at all, and not pleasant to live through, but the kind of thing you could recover from at home. 15% of the cases were moderate and needed some kind of medical support, such as supplementary oxygen, to get through. And 5% of the cases were severe — these were the cases that needed a lot of medical care, some quite substantial medical intervention such as intensive care, being put on a ventilator, etc. to get through and survive. These numbers were available since mid February, when we were all discussing them on online forums, so anyone acting like this is a sudden surprise based on data from Italy is lying to you, but unfortunately Italy is also corroborating this data right now.
And this is really where I’m afraid the terrible nature of this disease arise. Let’s ignore the confirmed case bias for a moment, the fact that hospitals are just seeing the worst of the outbreak.
If 5% of cases require some kind of intensive care, then you can do the calculation. Perhaps 60% of people will get sick in the UK. Perhaps 5% of those need intensive care. That’s 2.1 million people needing intensive care at some point over the next however-many months. The UK currently has 1,000 intensive care beds available — 4,000 in total, but 3,000 of them are usually being used by people who need intensive care for other reasons, and of course, if they don’t get that intensive care then they will die. Sadly that’s not going to stop just because an epidemic is on. That’s the mismatch we’re talking about here. 2.1 million vs 1,000. So you can see, unfortunately, in an epidemic like this one… it almost doesn’t matter how good your healthcare system is. If these numbers are anything like correct, if you happen to need care at some point, you are unlikely to be able to get it. And I think it is this prospect more than anything else that has lead to the apparently draconian lockdowns that are taking place now across much of Europe, affected areas in the United States, and which have been seen in places like China and East Asia for a while already. Even if these numbers are off by a factor of 10, even if we can then get 5x as many beds ready, you could still see tens of thousands or hundreds of thousands of people dying and in need of ICU in my country. It’s that serious, which is why the shutdowns and the lockdowns, as irritating as they might be, are completely justified in my view.
They have also observed the following as far as fatality rates go. Now I should say this is specifically for the outbreak in Wuhan. For part of this outbreak, the healthcare system was overwhelmed and people were not getting the ideal treatment. So we have a tricky situation. Do you quote the fatality rate for the whole disease, as best as you can estimate it, assuming everyone gets great medical care? Or do you take into account the likely strain on medical care systems, which will cause fatality rates to rise? I hope I’ve persuaded you that assuming medical care will be severely strained is a good assumption right now, so Wuhan numbers might be the best we can hope for. Here they are:
One interesting thing here is that they tried to account for the two effects that we’ve talked about already — the bias in hospitals not identifying people with symptoms, and the delay bias that arrives because it takes a while for the cases that are going to die to, unfortunately go on to die.
You’ll see one obvious thing here: this disease is much much less severe in the young than in the elderly, although deaths amongst young people are not unheard of.
There is also data surrounding people with underlying health conditions that suggests the death rate amongst people who are, say, diabetic, or with high blood pressure, or with cardiovascular or breathing issues, could be 6–7%. I can’t set too much store by this data myself simply because it’s also correlated with age. As you get older, you are more likely to develop these underlying health conditions. For this reason, it’s really difficult to say that any given diabetic has a 6–7% chance of dying if they get coronavirus, because a lot of those people will be older. At least I hope this is the case. My brother is diabetic, although he’s 21. So neither the 0.2% rate for people in their 20s or the 7% rate for diabetics applies to him. But both of those numbers are much too high for me to be happy about the risk over the coming weeks and months.
We also know from this evidence that the fatality rate is approximately twice as much in men than it is in women. I say to my female friends a lot that having a better immune system is basically a small repayment for the amount of nonsense they usually have to put up with on a day-to-day basis, and they tend to agree with me, so there we go.
Okay. Let’s talk about the good news a little bit, because I think there is potentially some good news on the horizon that is not yet widely circulated, although again, it all comes with a grain of salt, because unfortunately, pandemics are not replete with good news; they really aren’t.
The first is to go back to that case confirmation bias. People at the hospitals are only seeing the worst cases; they are only seeing the tip of the iceberg; they are only likely to see people who are the worst affected. It’s actually quite rare that a full population is entirely tested for coronavirus, because testing is rightly prioritised to people with bad symptoms. But when this has happened, there have been some interesting results.
The Diamond Princess cruise ship, as I’m sure you’re aware, had a single case of coronavirus that eventually resulted in hundreds of cases due to the botched quarantine attempt on that cruise ship. They tested everyone on the cruise ship; it’s a rare example of a full population being tested. They found that half of the patients on there had no symptoms at all from coronavirus when they tested positive. Some of them may have gone on to develop symptoms, but we don’t have that information. It’s worth noting that as of today’s best information, 696 people on the Diamond Princess got coronavirus and 7 died. That’s around 1%. 15 people are still in critical condition, although this was higher a few weeks ago, and 450 are already listed as fully recovered. It’s worth noting that the Diamond Princess was a cruise ship, so the average age there is pretty high compared to a normal population, just because of how cruises are! Similarly, a town in Italy recently tested its entire population for coronavirus of around 3,000 people. The article was in Italian, and certainly isn’t confirmed just yet and certainly isn’t a scientific paper, but it also suggested that up to 50% of those tested had no symptoms. Iceland, a small nation of 300,000 people, recently tested a big sample of around 3000 people at random. They found that around 1% of the country had coronavirus on average — so perhaps 3,000 people have coronavirus in Iceland. Half of those had no symptoms at all; half of those had symptoms that resembled a mild cold or flu. So far, as of today, there has been just 1 death in Iceland and only 1 person is listed in critical condition, although of course these numbers can change.
Until we have large-scale screening, we can’t really know what fraction of people have a mild illness and what fraction of people have no symptoms at all. These are only anecdotal lines of evidence so far. And in some ways, this is of course a mixed blessing. The fact that so many people can be infected with this virus and apparently not even notice they have it is great. If you’re like me, you think — this is a coin flip; perhaps I already had the virus and never noticed it; perhaps, for me personally, not caring about anyone else, the epidemic is already over.
And it’s good for those figures that we talked about earlier, too. You can cut the death rate in half. You can cut the serious cases rate in half. You can cut the hospital treatment cases in half. If this is borne out by further studies, and I… I really just hope against hope that it is — then we are talking about millions of lives here that are potentially going to be saved. That’s not a small thing, it really isn’t. And who knows — maybe we’ll get even luckier, and there’s even more serious case recognition bias than we currently expect, and it will turn out that the death rates can be reasonably low ultimately.
But of course the sting in the tail is that if as many as 50% of people really are completely asymptomatic, then this thing is going to be impossible to contain, which is what we’re finding right now. It would allow them to run around and transmit the disease without even knowing it — and remember, no one is responsible here except the virus, which is just a dumb strand of genetic code that has no conception of morals anyway. Given that this thing is out there, and it can’t be contained now, this is likely the best news we could get for a while. We don’t know if it’s true yet. Which again is why I completely support the lockdown which frankly should have happened earlier, and if it hasn’t happened in your country then it really should now if at all possible.
I saw one fascinating study that suggested — using a mathematical model, which is again imprecise and a really big unknown factor here — that as many as 5% of the population of Wuhan could have the virus. If that’s true, then the 3,000 deaths there would mean that even with a strained medical system, the mortality rate of this epidemic was just 0.5%. I don’t know how true that will ultimately prove to be, and I don’t know how strained our medical systems will get and how many people will die for other reasons during this epidemic, just as I can’t be sure what R0 really is, but if this is true, then it’s a sign that the worst-case scenarios I posited earlier might not actually be correct at the end of the day.
Luckily, some new information should be arising pretty soon about how much this confirmed case bias is really affecting our thinking at the moment. This is in the form of a serology test — basically, they test your blood for antibodies. The body’s immune system is really a remarkable thing. When it detects a new pathogen, a new virus or bacteria, it “learns” how to produce proteins that destroy that virus or bacteria. This is what gives you immunity to diseases — these antibodies, these proteins that your body produces to repel the invader, stay in your system for a long time. So with this test, we’ll be able to see how many people have really contracted the virus — even if they just had a really mild case of it — and then we’ll get a proper estimate of the actual fatality rate and the actual severe case rate, and with any luck, they will be better than the 1% fatality and 5% severe case rate with good healthcare that we’ve seen so far, and that’s really the best that we can hope for at the moment. It’s worth pointing out that a disease which spreads very very quickly and has a very very low fatality rate will look a lot like a disease that spreads more slowly but kills more people. So in some ways, now that this is out there and impossible to contain, that’s the best we can hope for.
So these are the three principles that keep coming back to me in this pandemic.
1. Delayed reaction; deaths lag behind symptoms which lag behind infections; you’re always further along the curve than you think you are.
2. Exponential growth limited by R0; the potential for this pandemic is ultimately really very great indeed, perhaps up to 60% of the population in total; growth is extremely rapid and a few weeks is all it takes to go from hundreds to thousands of cases, and then from tens of thousands to a million.
3. Confirmed case bias; the situation on the ground at hospitals will be very bad, but it might actually overestimate what’s going on because a lot of less severe cases are missed by testing and don’t end up in hospital at all, particularly asymptomatic cases.
But there is a very important dimension which has been missed so far, and that’s what we can do to help. Social distancing measures — this idea to keep away from people — are absolutely crucial right now, especially if we don’t know who does or doesn’t have the disease due to the high fraction of asymptomatic cases. It’s absolutely vital that we keep social contact to a minimum, work from home where we can, and so on. The economic disruption from this — which will probably have to be in place for many months — is going to be vast and all of our efforts should be focused on fighting the pandemic or fighting that economic disruption, in one way or another. Make no mistake here. This thing is going to define the year 2020, and perhaps 2021. It’s going to leave an impact on the following decade like a world war, or like a great depression would. We haven’t seen anything like this in our lifetimes; we are living through history; and it’s going to be a long-haul situation for all of these things, I’m afraid, as you have doubtless been realising already over the last few days. Even if social distancing is successful, if the numbers hold up, then this is going to be a truly awful situation for a long time to come. You can see this pretty simply if you go back to remembering the idea that perhaps a million people might need perhaps a thousand ICU beds in the UK. If social distancing and other measures to slow the spread can successfully slow the spread, then you must understand that we could need to do them for years and years for everyone who needs an ICU bed to get one — naively, let’s imagine that everyone only needs ICU for 1 week, and we can upgrade capacity to as many as 10,000 ICU beds nationwide, which would obviously be huge. Then it’s still two years to “safely” process all of those people (the death rate would still be somewhat high as ICU can’t save everyone, but it wouldn’t be four or five times higher as it might be if the ICU is not there for the people who need it). And this is assuming our social distancing measures can perfectly control how many people get the disease at any given time, which seems very naive to say the least.
Treatments. Viruses are notoriously difficult to treat. The main reason for this is how viruses reproduce versus how bacteria reproduce. Viruses are bizarre things, they really are; they can only reproduce inside a host cell; once they invade that, they convert it into machinery for creating more and more of themselves. So finding a chemical that will destroy just the virus, leaving the human cells that it lives within intact, are really difficult. Sometimes they try to block the virus from ever entering a cell; sometimes they try to muck about with the chemical processes that occur within the cell. Lots of the antivirals that do exist have some pretty nasty side-effects. They are closer to chemotherapy than antibiotics is what I’m trying to say here. Plenty of them are currently being trialled for use with coronavirus. There is some very early preliminary evidence that chloroquine, a drug used to combat malaria, and remdesivir, may be effective. But at this stage they are basically throwing everything they can at the problem to try and figure out what’s going to work here. I expect that only the most severe cases will ever try to use antiviral drugs, for all of the reasons that I’ve described. It’s going to be difficult to work this out, in terms of how effective they can be, and how much we want to use them, and on whom. In a few months, a new one for this might be manufactured.
Vaccines. Work on vaccines is already underway. With a vaccine, as you know, the aim is to introduce a little bit of the dead virus into someone’s system so that their immune system can recognise it and learn to produce antibodies without giving them a full-blown illness. But vaccines are very difficult to make. No vaccine for a coronavirus has yet been manufactured — they gave up on researching SARS after it became clear that this outbreak had ended in 2003. [Again, in the earlier pandemics episode… I feel like I was kind of naive about how easy it was to make vaccines, but I did suggest that it might be worth spending more than $500m on researching vaccines, because the consequences of a real pandemic would be much worse than that, and it’s necessary for society to be geared up to combat unlikely but massive threats. Where I’m sitting now, the stock market has collapsed by around 33%, wiping trillions of dollars of supposedly real money off the face of the market and very real money out of your pensions and savings accounts, and it seems like if only spending billions upon billions on vaccine research had been done in advance would speed up the end of this lockdown by a few short months, it would be unbelievably worthwhile.]
It seems likely that a vaccine will be manufactured, and it will be made. Trials are already underway. The issue is that you need to test, test, and test again. If a virus kills 0.5% of people who may not ever get it, but the vaccine kills 1% of people consistently, the vaccine is more dangerous than the virus. There are some vaccines, like the Ebola vaccine, where the disease you are protecting against is so often deadly that it might be “worth” a 1% risk of dying from the vaccine to avoid a much much higher risk of dying of Ebola, especially if you know that you are going to be exposed to the Ebola virus. In this case, they need to test this thing very thoroughly and observe if there are any medium-term consequences for the patients who get it. It’s this testing that takes the longest time, and so when all the experts and the pharmaceutical companies say that a vaccine could take a year or 18 months to fully develop, sadly, I do believe them. Even that would be a record. We’re talking about something, after all, that has never been done before in terms of manufacturing a vaccine against a coronavirus. So we can’t rely on that to save us in the short term.
So, after all that, what does this mean, do I think?
The long-term outcomes for the virus itself we can try to infer from past pandemics. Sometimes pandemics have faded in the summer only to return in the autumn, as with the Spanish Flu in 1918–1919. We know that coronaviruses are amongst the viruses that circulate amongst us and cause the common cold. It could be the case, that, in time, this new coronavirus will come back in a much less deadly form and will be just another one of the “endemic” diseases — i.e. circulating amongst us without causing massive outbreaks — that we have to deal with every single day. It could be, as is the case with some outbreaks, that the whole thing burns itself out and we don’t see this virus again. I think the idea that this is going to eventually circulate and get a lot of people until it becomes endemic and maybe mutates into a less dangerous form is probably the most likely outcome here. There’s more evolutionary pressure to select for less deadly than more deadly forms, for the reasons we discussed. It seems likely that people who get this strain will be immune to this strain, so we’d better hope that it is the most deadly strain of the new coronavirus that we see in the future. But without being a virologist, I don’t know for sure.
What does this mean for society in the short-term?
I’m afraid, very difficult things. Every major healthcare system in the world is likely to soon be overwhelmed. If you can do anything to support healthcare workers, now is the time, because they are going to be on the front line, saving lives, for the forseeable future. Individual societies will face difficult choices between draconian lockdowns which can suppress the epidemic — for a time, for as long as they can be maintained — and a spiralling epidemic which will result in many thousands or hundreds of thousands of dead. This will be the case for the next year or two at least, until a vaccine can be developed and widely deployed, I’m afraid. I can’t see any way that we get out of this pandemic without a substantial number of people dying. I really wish that I could, I really do. But the economic and societal shock is going to be huge; the shock to the medical system and to the most vulnerable populations amongst us is going to be huge; all this much seems, sadly, obvious now. I think we’re going to see a lot of very interesting social phenomena. I think we’re going to see how people and populations cope with stress and trauma that’s unlike what they normally have to deal with. We might see strange things — young people getting the virus intentionally, for example, so that they can join the ranks of the immune and return to some kind of normal life where they don’t have to worry about getting sick and they don’t have to worry about infecting the people around them, as examples. We’re going to see a lot of boredom and quite a lot of tragedy and hopefully quite a lot of stories of human resilience, because we’re going to get through this as a species, just like we’ve survived everything else up until now, but it’s going to be a bumpy road.
What does this mean for society in the long-term?
I don’t know. I’m not foolish enough to pretend I know these things. There’s a famous quote by a historian about the French Revolution — asked about what they thought the impact of the French Revolution will ultimately be. “Ask me in a few hundred years when it’s over, was the response.”
I can tell you what I hope will happen. Those who listen to the show will know how long I spent talking about existential risks, talking about nuclear war, talking about pandemics and bioterror and climate change and everything else that’s so concerning. And when I focused on these apocalyptic scenarios, these catastrophic risks, a lot of people essentially told me to stop being so morbid. I understand that. I understand how hard it is to think about these things. Focusing on the worst case scenario for so long — in climate change, where I do my main research, and now for this pandemic — has not been even remotely easy. I can understand why people don’t want to think about it and bury their heads in the sand and not even deign to prepare for the worst-case scenario. But it’s here now, and that failure to prepare is going to look increasingly terrible.
I want to get personal here — I will get personal here. I had to stop the show in November and it’s only coming back now in March. I was sick. I don’t know what the cause was — maybe it was some viral infection, which is what the doctors reckoned — maybe it was depression; I honestly cannot tell you. All I know is that for about a month I was pretty much unable to think properly, or get out of bed most days, or form coherent sentences in conversation most days, and towards the end of it I was wondering if I would ever get better, or if I’d just be stuck with whatever this debilitating condition was. And thankfully I did get better; I did recover — not totally but I’m on the mend. But it made me re-evaluate a lot of things. It actually made me more concerned about this pandemic. And I wrote about it in my diary a few weeks back, the following.
11th Feb, 2020.
This feels stupid to write, but it might make me feel better. That’s the only reason I’ve ever kept a diary, so why break the habit?
I was pretty sick last year — ill enough that I was confined to bed, feeling weak and fatigued, for several weeks. I still don’t know what was wrong.
Now there is a novel coronavirus — something that came out of the wet markets in China and has already killed a thousand people there. It seems to have all of the characteristics of the deadly viruses I studied in the Pandemics TEOTWAWKI episodes — spreads without symptoms, can kill up to 2% of those infected (maybe more if the healthcare system gets overwhelmed). Eight people in the UK already have it… and people don’t seem to understand why I’m really nervous and worried about this thing taking off here.
I don’t know what’s going to happen, obviously. At this juncture, it’s hard to say for certain — so few of the facts are well pinned-down, and there is always an element of luck to all of these things. Plenty of people — maybe more than we know — get this illness and essentially don’t experience any symptoms worse than an ordinary flu (it’s the ones that do deteriorate that you have to worry about…)
At the same time. I am struck, as I was when researching these TEOTWAWKI episodes so many months ago, by the fact that we really take an awful lot of things for granted. We take the stability of the systems — of the world around us — for granted. We take the idea that each day is just another day to get through, of no particular significance in the grander scheme of the narratives of our lives. We will casually plan things weeks, months, years ahead — as I’m doing right now, my calendar is full of people to interview for my job in Parliament — as if we have some level of certainty that we won’t be suddenly struck by an unexpected, and rapid, change of priorities. As if we *know*. As if we’re blessed! And yet, coronavirus or no coronavirus, extinction-level meteorite event or no extinction-level meteorite event, these certainties are so often shattered. “I’ve seen this happen in other people’s lives and now, now, now it’s happening in mine.” And whole societies, too, see their certainties crumble around them… it happens all the time. Of course, it’s like mortality itself. You can’t ever think about it — you can’t allow yourself to think about the inevitability of that extraction from everything you know, and the impossibility of fathoming what — aside from nothingness, and what is endless nothingness — would come next if it were to happen to you.
I fucking hope that all the plans and the half-formed, half-baked notions about the distant future that were in my head can all come to fruition, one way or another, and I can go back to worrying about the years and not the weeks and months. I really do. It’s a luxury — like good health — that most of us take for granted when we have it, and rue the fucking days we took it for granted when we no longer can.” That was the 11th February, 2020, when eight cases were confirmed in the UK.
I don’t say this to say I told you so. This is because I’m not that smart. I’m an idiot. I’m not an epidemiologist or a virologist. They knew that something like this would happen at some point and they’ve been warning us for years. There have been people warning that we are not ready for the next pandemic for as long as we can remember. (Bill Gates, notably, back in 2015, but genuinely for as long as anyone can remember.) These epidemiologists and virologists knew that this virus was this dangerous weeks ago and they’ve been warning us for months, they’ve been warning governments for months that more action was going to be needed. If your government now appears unprepared or is only now taking actions that would have been useful earlier, like constructing extra hospital capacity, testing people, buying masks, or building more ICU units, you have to ask yourself why those warnings were ignored or downplayed early on. It took China to lock down one of its biggest cities for me to realise that this was a problem, at which point it should have been clear to everyone. So I’m not a prophet at all. I’m far less smart than these people are. The point is that we were warned for so, so, long and we ignored the warning. SARS was 2003; MERS was 2011; this happens approximately once every ten years that a virus crosses over in this way; how many other epidemics have there been? How many other times did we play Russian Roulette until one of these viruses was sufficiently bad enough that we couldn’t stop it?
The fact that we as a species have spent more money over the last few years on nonsense, on pointless financial speculation on tech companies with flaky business models, on skyrocketing prices for essentials like real-estate, on making the uber-wealthy uber-wealthier, on halloween pet costumes or luxury holidays or luxury food or whatever other trivial and irrelevant nonsense looks so trivial and irrelevant now; it has come back to bite us in the bum in a very big way. This was not unforseeable. It was forseen; just like all the other catastrophic risks we covered way back in 2017.
So I don’t know what will happen in the long-term. But I hope it will inspire a change in priorities. Right now, your priorities are shifting from whatever they were a few weeks ago to looking after those closest to you and your own health. Society’s priorities must also shift as a result of this. We have to take these low-probability, high-risk events more seriously. We have to do more planning for forseeable catastrophes. We have to be smarter. We have to cooperate internationally more — just like the scientists, in the wake of the nuclear bomb, concluded that with such deadly power on the Earth there could only be “one world or none”, we have to find a way to get closer to one world so that we can avoid the threats that are common to us all. This virus — climate change, nuclear war, existential risks in general — they affect everyone to some extent or other. We need to be smarter, more organised, willing to devote a greater fraction of resources to trying to avert these catastrophes. Willing to accept that the party might end, that things might shift, that there might be downsides, and that we might need to be prepared for things to get worse. Willing to accept that it’s worth apparently “wasting” resources on preparedness for these low-probability, high-consequence events, because those resources won’t look wasted when you’re staring down the barrel of cure, having failed at prevention: they really won’t.
Who is to say whether it would even be successful? Who knows if any system that we can construct would be robust enough, clever enough, wise enough, and possessed of enough foresight to deal with something like this? But at least if we tried, we would not have to deal with the massive, massive downside that arises when something goes wrong. This is going to expose all of the flaws in our existing systems of running the world, in the existing fabric of our society, in the priorities that we have had before, and you know what those priorities were, and why they are failing us now, and ways in which they will continue to fail us. But it is the case that disasters offer us an opportunity to reshape society — I was actually working on a podcast series to that effect which is now going to take on a whole new light — and perhaps, perhaps, when this is all over, we will have a chance to build a much better world; a wiser world; a safer world.
What can we do practically? Mutual aid groups right now are being established across the world to help people who are medically vulnerable to this illness through age or pre-existing conditions. You can help them if you are young and healthy by getting shopping for them, by checking in on them and making sure they’re okay, by renewing the ties with people you had neglected, by doing whatever you can to support the healthcare and emergency services in their hour of need, and by social distancing to slow down the spread of this pandemic as much as possible, you will protect people. A few weeks ago, when it was a much smaller issue than it seems in Europe today, I bought up some fancy masks to protect myself from this. They’re not necessary. I am young and mostly pretty healthy and will make it through. I gave them to the A+E at my local hospital earlier this week, where they were happy to have them. Again, it was the very least I could do. If you have that kind of resource at your disposal, please consider doing the same.
What can we do in general?
We will get through this; there is no doubt about that. The human race has incredible fortitude — you will find that you yourself are capable of dealing with things that were unimaginable just a few weeks ago. We will survive — indeed, as I mentioned earlier, for most of us, the chances of a serious outcome here are very low indeed. Society will emerge, blinking and dazed, from quarantine or pandemic. But there are some things that are required to help us do so. It is more necessary than ever to trust science and reason — but not to allow our knowledge and our calculations to get in the way of our compassion. It is more necessary than ever to focus on people’s lives over profit or anything else short-term. It is more necessary than ever to reach out to your fellow human beings and do what you can to help them. It is more necessary than ever to be brave, and to be kind. That’s all we can do.
That’s all I have to say. If people want a Q+A asking my opinions, whoever ends up listening to this, or more pandemic content and they let me know, or if something major changes that I feel really needs an update, I am going to produce more shows on the subject. Otherwise I’m going to return to the stuff I had planned before this all kicked off and hopefully give you all something distracting to listen to over the coming weeks and months. Please take care of yourselves and the people around you.