More Disease, Less Diary

Wednesday 1–Saturday 4 April 2020

This coronavirus, it’s not a bloody black swan. A black swan isn’t a known unknown, and we knew a plague would come—we just didn’t know when. If it were a meteorite, that wouldn’t be a black swan. If it were a flood, or a fire, or a nuclear war, or a cybernetic attack, it wouldn’t be a black swan. A black swan is an unknown unknown. If I can think of it, if I can write it down, it isn’t a black swan. Nassim Taleb called them ‘black swans’ because, apparently, in Europe people used to say, 'That’s about as likely as a black swan'. Today, walking around the Mona waterfront I saw black swans. But to Europeans before the Australian invasion, black swans weren’t a known unknown. They weren’t on the radar. A viral pandemic, that was about as likely as a white swan. There are Hollywood movies about such things, for Christ's sake.

So don't tell me this disease is a black swan. It's a black panther. It creeps up—slowly, even elegantly, and then quickly, viciously, powerfully. But black panthers aren't really that dangerous . . . It's a blacklist. If you're on the list, through no fault of your own, you're isolated. And if you even know someone on the list, you're out of work. Or it's a black bear, too big to take on, too fast to outrun. But bears sleep through the winter. It's worrying, for us antipodeans, that the beast we are facing might get grumpier in the winter.

But bears and panthers and swans won’t wear face masks. The thing I'm finding hardest to understand about WHO and government recommendations in relation to COVID-19 is the opposition to community face masks. It's noticeable that countries that are performing well in this crisis (mostly Asian) are wearing face masks. WHO is recommending against them because: cheap ones don’t work perfectly (and there aren’t enough good ones for medical professionals); and they say people touch their faces more often when they are wearing masks, and thus might carry infections to their mouths.

To me, they are missing the point. Every time I've been to an Asian country, I've seen a few people wearing masks. They are not trying to protect themselves from me, they are trying to protect me from them—they are wearing masks because they have a cold or the flu, and might be infectious. Now, it may be that cheap masks don't protect the wearer, but they do protect the community. And it doesn't matter if they don't work well, provided they work at all. If people wear masks in addition to social distancing, asymptomatic carriers are less likely to infect others. WHO and many governments show why individuals aren't served by masks, but they don’t address the fact that the community is served by a mass population of wearers. Masks serve the Precautionary Principle.

However, just because Asians wear masks, and Asian countries seem to be controlling this coronavirus effectively, that doesn't mean that masks work. It may be, for example, that conforming societies wear masks, and conforming societies also maintain social distancing and stay at home. As we hear often, and ignore nearly as often, correlation does not prove causation. However, correlation suggests causation is possible, so the defensive strategy (the Precautionary Principle) is to react as though causation is demonstrated.

Another potential contributory factor to the success of Asia in this battle is the recent use of the tuberculosis vaccine BCG. BCG has been widely used in Asia, including Japan, which is showing a lower exponential multiple than western countries despite not closing everything down (to my limitless satisfaction, Japan still has horseracing). An interesting classifier for BCG coronavirus efficacy is Germany; the former East Germany used BCG, and it is expressing a lower growth rate than the rest of Germany. That's probably a fluke, and there could be other explanations, but it's justly being investigated. Australia used BCG until the '80s, and it's still being used in Aboriginal communities (a lovely, but unlikely, consequence might be Aboriginals enjoying better health outcomes than their white cohort—that doesn’t happen often).

Australia's mortality rate is low (as at Saturday morning, 4 April, 28 deaths from 5462 diagnosed cases; 0.5 per cent), and that might need an explanation. High exponential growth conceals deaths, because there is a rapidly expanding population of people who have just contracted the disease, but haven't yet had a chance to die—that's not Australia, though. Our case rate growth has levelled off for a week, with no consequent increase in mortality rate (however, secondary mortality from cruise ship passengers will hit over the next few days). The US looks to be in that situation now (lunchtime 3 April, US diagnosed cases 245,070, deaths 5949, mortality rate 2.4 per cent; compare South Korea with the most resolved cases mortality rate of 1.7 per cent), though its death rate is not that low. The US is grappling with another factor that increases the apparent mortality rate: a low rate of testing. Deaths self-identify, so if there are more unidentified COVID-19 people in the population, then the fatality rate looks higher. The upshot is the US mortality rate is uninterpretable. (Later: 5 April, the US has ramped up its testing, but despite misleading claims from President Trump, its rate is still around one third of that of Australia—and there are now 308,533 diagnosed cases, 8376 deaths, a mortality rate 2.7 per cent).

The UK's situation is also diabolical, but at least they publish decent statistics. As of Wednesday 3 April (Aus time), the UK had conducted 163,194 tests, with 33,718 of those tests positive. Surprisingly, Australia has conducted more tests (214,000), even though it has only 37 per cent of the UK population. The UK and the US may be paying for an ideological resistance to the removal of personal freedoms, but testing isn't an ideological problem. Australia's tests resulted in a positive rate of 2.5 per cent compared to the UK's 21 per cent. Australia has a large problem, but that statistic shows the UK is properly in the shit. It could also mean that Australia is poorly targeting the coronavirus reservoir with its testing, however its testing targets seem to be a superset of those of the UK. False negatives also need to be accounted—tests of coronavirus-infected people which show no infection. If Australia's false negative rate is much worse than the UK's, that would indicate that both countries are wallowing in crap.

Testing bias is probably the explanation for Germany's relatively low mortality rate. Germany is apparently testing 500,000 people a week (or is about to), so it'll be detecting asymptomatic people. Here's a mortality summary: countries with exponentially expanding cases show lower mortality rates but they are often the countries with medical systems under pressure, which increases actual mortality. Countries with wider testing show lower mortality because they catch more asymptomatic people.

Other factors that may affect mortality rate: smoking incidence (Italians smoke about 50 per cent more cigarettes per capita than Aussies); age distribution (23 per cent of Italians are over 65, in Australia, that's 15 per cent; but in Japan, it's 28 per cent); age at infection (in Australia, younger people are contracting COVID-19 because most infections, so far, are travel related—however, many of the fatalities are older people from cruise ships); greetings (Europeans 'kiss' when greeting, Asians don't, Australians don't much); urbanisation and urban population densities (Australian cities are big, but people live in houses); and luck. Spurts of local exponential growth are extremely dependent on initial conditions—each infected person infects an average of about 2.2 people (before intervention) but there is large variation, so it could be none or ten. After a few weeks, a pocket of potential infection might contain no coronavirus-positive people or it might contain 250.

To understand the mortality rate, some insight is needed into asymptomatic cases, particularly since most localities are testing only people who display symptoms. The first port of call in this investigation is the Diamond Princess. Japanese authorities tested 3063 of the 3711 passengers, and detected 634 infections, of which 328 were asymptomatic at the time of testing. Some of those will have gone on to develop symptoms. A paper applying a credible methodology¹ suggests that 48 per cent of infections would have remained asymptomatic (although others interpret this paper differently). Cruise ship passengers are older than the general population (much older—on the Diamond Princess, the average age was 58, my age). The paper recalibrated Chinese data based on the Diamond Princess information (with appropriate mortality by age for the Chinese population). According to the study's calculations, the modified mortality rate in China is about 0.5 per cent. And asymptomatic infections constitute about 55 per cent of all cases. The evidence suggests about half of the infections are asymptomatic. And, of course, those are the ones not typically discovered with testing. Another set of data comes from the Italian town of Vò. Everyone in the town of 3300 was tested, and about a half were asymptomatic (some may later have developed symptoms).

The combination of asymptomatic carriers, limited testing and false negatives suggests that a reservoir of infection will remain in a population unless isolation is total, and thus those models suggest that a reduction in social distancing will create another explosion in the number of cases. I guess that's why the Australian authorities are saying the only exit strategies are a vaccine or a cure.

In general, I think our elected leaders are doing okay. Judging the quality of decision-making by the quality of outcomes is fraught, and it'd be nice to have a few less docked cruise ships and a few more masks, but Scott Morrison, federally, and Peter Gutwein (Premier of Tasmania, where I reside), seem to be communicating appropriately, and following the Precautionary Principle. On Thursday, Premier Gutwein barred prostitution, to which my friend Emily commented, 'Wives will have to be upskilled.'

And I'm also doing okay. I've got the deserted Mona grounds to wander around in, and I guess I could visit the museum. Sunday (daughter) hasn’t seen another kid for sixteen days—she just did a dance class online. Yesterday, she compered a cornbread-baking and ladybug-painting video (with her mum directing). Grace (another daughter), in Sydney, has maintained her sanity with the acquisition of a new puppy. We won’t see each other for months. When this is over (if this is ever over), we might return to a different world. Right now, I think I'll be seeking the company of others. I was happy to stay home, until I had no choice but to stay at home. On Solaria, the planet on which Isaac Asimov's The Naked Sun is set, no-one sees anyone except to have sex. We might go that way, and we might not. Forecasting disease outcomes is tough but forecasting social responses to those outcomes is a recipe for disaster. Everyone might be different. Or everyone might be the same. Communities might bind more tightly, globalism might give way to localism. China might thrive, the US might flounder (or vice versa). Nations may endure civil wars, and human rights might languish under residual police states. We might return to an expanded ideological divide, or we might see the emergence of a pragmatic pluralism. We might all get together and appropriately respond to other existential problems, like global warming (while singing 'Kumbaya'). Or we might give up hope. In Australia, proponents of maintaining the augmented dole might prevail, or expanded public debt might justify a tightening of the fiscal reins. Mona might reopen with renewed novelty and vigour, or to a distracted public with a Netflix addiction. Wealth may be redistributed, or undermined. Prostitutes may be unable to compete with upskilled wives. I can't wait to see. But I have to.

Some Chinese say, 'Better to be a puppy during peacetime, than a person in times of war.' (They do not say, 'May you live in interesting times.') Despite the immense suffering of others, I'm at peace in this time of war. But I'm a dog with a bone.

1 Timothy W Russell, Joel Hellewell, Christopher I Jarvis, Kevin Van Zandvoort, Sam Abbott, Ruwan Ratnayake, Centre for the Mathematical Modelling of Infectious Diseases COVID-19 working group, Stefan Flasche, Rosalind M Eggo, W John Edmunds, Adam J Kucharski, ‘Estimating the infection and case fatality ratio for coronavirus disease (COVID-19) using age-adjusted data from the outbreak on the Diamond Princess cruise ship, February 2020’, Eurosurveillance, Vol 25 Iss 12, 26 March 2020.