25/03/2020

Research / Other publications

Waiting for Coronavirus

How strong is the case for mitigation?

Budapest has had exceptionally good weather the last fortnight until yesterday, when it started snowing. Since spring was already arriving in defiance of the calendar, this has produced the oddity of snow on the blossoms. That’s not a bad description of our mood in recent days, which is one of melancholy delight in the beauty of the restored Budapest as we strolled in ones and twos through the quiet streets and almost deserted parks last week, and as we pace around our apartments today waiting for the virus to strike us.

It has already struck some of us. The latest figures for Hungary are 167 confirmed cases of infection and seven deaths. All but one were elderly and most had unrelated health problems. But the general expectation is that these figures will rise sharply in coming days in imitation of Italy’s long crisis. Without that expectation, the figures for Central Europe would not look so alarming, especially in comparison with other regions. Here are the figures for infections and deaths in neighboring countries: the Czech Republic, 1,165/1; Poland 684/8; Romania, 576/5; Croatia, 306/1; Slovakia, 186/0; Ukraine, 73/3; Belarus, 81/0; Moldova, 94/1; and Slovenia, 414/3. The Baltic states are doing slightly better; most of the Balkans, including Greece, slightly worse, but all are in the same ballpark.

Consider now the same figures for Western Europe: Italy, 59,138/5,476; Spain, 33,089/2,206; Germany, 26,220/111; France 16,720/674; Switzerland, 8,547/118; the United Kingdom, 5,748/290; and the Netherlands, 4,217/180. For comparison: the U.S. has 35,241 infections and 473 deaths; Russia, 438/1; and China, where the Coronavirus broke out in November, 81,496/3,274. (These figures are taken from the Johns Hopkins Coronavirus Resource Center tracking map and updated continually. I’ve given you the figures for 10.00 a.m. EST on Monday so that you can check them against today’s to see the different rates at which the virus spreads.)

Do these various figures justify the prevailing broad narrative, from both officialdom and the media, that unless we all adopt a total lockdown as soon as possible, all the above countries are doomed to follow the Italian path to mass infections overwhelming health services and producing hundreds of thousands of needless deaths? I don’t think this narrative can be dismissed lightly, but it should be examined skeptically, because there are some distinct oddities in the statistics on which it’s based.

Why, for instance, are the figures for deaths in Italy and Germany so much at variance when their figures for infections differ only moderately? The latter can be seen as different points on the same curve; the former need some additional explanation. What might that be? Are Germany’s health services superior to Italy’s? Given German efficiency, that’s not an unreasonable guess. But the experts seem to agree that health services in Lombardy, where the infection took hold, are among the best in Europe. Quite simply, good health services were swamped by the number of infected Italians needing scarce and costly treatments. Are border controls tougher in Germany? No. Before the spread of COVID-19 took serious hold and prompted caution, the two countries had the same EU open borders penetrated by the same migrants. Do age differences explain some of the divergence between the two death rates? That’s more complicated. Germany and Italy both have older populations hovering around a median age of 46. Germany is very slightly older, but the Italian population is less healthy — it smokes more heavily and is more resistant to antibiotics. Additionally, the population of those Germans infected by coronavirus is younger than the same group among Italians. The virus first arrived in Germany via healthy, young, and more-resistant people enjoying themselves at ski resorts in Bavaria and northern Italy. They are less likely to die of coronavirus infection than are Italian patients, whose average age is 67. That may change as the virus spreads to other age groups, but for the moment it seems to bias the statistics slightly in Germany’s favor. Not enough to explain, however, a divergence of the magnitude from the statistics from Italy.

The decisive factor seems to be the different ways in which statistics of patients who die from the virus are recorded. When this is investigated, Italy’s figures look unduly high not only against Germany’s but also against those of most other countries listed above. A thorough examination of this appears in “Global Covid-19 Case Fatality Rates,” by Jason Oke and Carl Heneghan of the Centre for Evidence-Based Medicine (CEBM) at Oxford University, This study is not without its technicalities, but it is overall a fascinating and readable analysis, of which I can give only a brief taste here. It points out that case fatality rates (CFRs) vary over time, tend to be higher in the early stages of the epidemic than later, and are subject to several kinds of bias. Italy is a prime example of this. As they write, referring to a report by Sarah Newby at the New Zealand website Stuff:

In Italy, all those who die in hospitals with Coronavirus will be included in the death numbers. In the article, Professor Walter Ricciardi, Scientific Adviser to Italy’s Minister of Health, reports, “On re-evaluation by the National Institute of Health, only 12 per cent of death certificates have shown a direct causality from coronavirus, while 88 per cent of patients who have died have at least one pre-morbidity — many had two or three.”

Recording the numbers of those who die with Coronavirus will inflate the CFR as opposed to those that died from Coronavirus, which will reduce the CFR.

And, indeed, in an Italian study of those who died, only three patients (0.8 per cent) had no pre-existing health condition, 99 percent had at least one preexisting condition, half had three, and the average of all was 79.6 years. Those figures give some indication of the difference between dying withcoronavirus and dying of coronavirus. That’s a somewhat technical difference to the deceased and his family, but it’s a significant difference to any judgment of how big a threat the virus poses to our peoples. Ignoring it inflates the total of deaths owing to the virus; correcting it must mean reducing the raw data of Italian CFRs to something nearer the German ones. As the CEBD authors point out, moreover, the various estimates of case fatalities all decline substantially from the start of an epidemic to its final tally. Swine-flu estimates fell fivefold over that period.

Reports in the past few days, moreover, suggest that over the weekend the Italian curve “peaked” and should now gradually decline until it dissipates entirely. That may take some time, and I leave it to the epidemiologists and statisticians to guess where both will end up. At this “peak point,” however, the Italian figures are 69,176 infected and 6,820 deaths, while Germany at an earlier point on the curve is looking at 31,260 infected and 156 deaths. These two countries point to quite different estimates of final fatalities.

Let’s look at a slightly more average case. The U.K. has 8,164 infections and 423 deaths. It has far fewer infection cases than Germany does and three times the number of deaths. And in the report of the Imperial College scientists — the one that underpins the policy of the British government — their estimate is that the coronavirus could lead to the deaths of between 20,000 and 500,000 people, depending on whether nothing is done or a quite draconian “lockdown” is imposed. And Boris Johnson has just done the latter. (Their estimate for the U.S. is between 500,000 and 2,200,000 deaths.)

Their report, though not without its technicalities, is closely argued and readable. My advice is to read it here. I shall not attempt to paraphrase it, since Peter Smith, a former colleague at Quadrant and a fine economist, has done a first-class job of both summarizing and critically analyzing it. But I will make a few points that strike me as relevant and important:

  • The IC report itself is a balanced and flexible document. Though its authors chose a policy of “suppression” over “mitigation,” that was a matter of emphasis rather than of a strict division. Most of the practical policies to tackle the pandemic — case isolation, voluntary home quarantine, social distancing of the elderly, social distancing of the entire population, closure of schools and universities — are proposed under both headings but in different combinations, timescales, and so on.
  • The report itself was not a departure from previous government policy — the “U-turn” much touted in the media — but the evolution of policy that was a response in real time to the dimensions of the threat posed by the virus. What led the scientists to propose a move from mitigation to suppression was data from Italy showing numbers of infected people so high that the hospitals were overwhelmed.
  • Without a policy of suppression — i.e., immediately halting the spread of the virus by quarantining the population — the National Health Service would be overwhelmed, as in Italy. But the virus would remain in the population and resume spreading when quarantine was eased. There would be a second upsurge of infections and deaths in the fall, as has happened in earlier epidemics and may now be happening in China. And so a second lockdown. Or third.
  • At the same time, a “lockdown,” partial or total, under suppression would gravely damage the economy, perhaps reducing the gross domestic product by a fifth to a third and creating mass unemployment. It would also halt the gradual expansion of “herd immunity,” which under mitigation would have meant that most people would have suffered the mild symptoms of a weakening virus, thereby becoming immune, while the elderly and at-risk groups were protected by quarantine measures reserved for them until the virus had been more or less eradicated.
  • Great depression or the breakdown of the nation’s health service? It’s not an easy choice, and to be fair to the scientists, they recognized this in the report, acknowledging both that there were crucial social and economic aspects of the crisis and that they could advise only on its medical aspects.

Once the report was released, however, all restraints of practicality were released too. A mass public pandemic of panic took over. The merits of suppression versus mitigation seem to me to lie narrowly on the mitigation side of the argument. But its great failing is that allowing a virus to spread, albeit to sections of the population resistant to it (while protecting the vulnerable), is a very hard sell. That failing was magnified by media that scented government incompetence in the mythical U-turn (a reversal denied by the IC chairman) and set off to prove it. And that press campaign was made as toxic as the coronavirus by the fact that large numbers of pundits, including some conservatives, are in the grip of a wildly irrational “Boris Derangement Syndrome” that leads them to believe the most exaggerated (albeit contradictory) charges against him — “He’s a fascist who wants to control everyone.” / “He’s a libertarian who won’t impose the necessary controls on people.” (Much the same syndrome can be seen in the United States; indeed, some say it was invented there.)

One result is that public opinion has demanded — and governments have yielded to — the imposition of lockdowns that go much farther than the IC report proposed. The report was, for instance, ambivalent-leaning-to-hostile with respect to school closures. They would not reduce transmission of the virus between children, since they would still be mixing outside; if they had been infected, they might transmit the virus to vulnerable grandparents given the task of looking after them by harassed parents; and those parents working in the health sector and emergency sectors, now greatly needed to deal with higher patient loads, would be kept at home. But governments all over Europe — except, to its credit, the Dutch government — have now closed schools in response to public pressure, even though the ministers in them will tell you privately they think this is a bad idea with likely bad consequences. Boris Johnson’s Tory government has gone to extreme lengths in this regard, literally imprisoning people in their homes, with only an hour or so for exercise or shopping every day.

It offers only a little solace that Boris is doing this reluctantly. It is still leading to the deliberate economic ruin of the country. President Trump sees the same thing and so proposes ending or avoiding a shutdown, but without a plausible way of dealing with the threat of the coronavirus in the longer term. There is a basic flaw to this approach. As Peter Smith writes: “The policies being adopted by governments are not tenable. They will bring about unquantifiable and crippling economic and social (and quite possibly serious health) consequences. Make no mistake, governments will be forced to reverse course . . . and adopt a different strategy.”

What he has in mind is a weaponized version of the mitigation strategy.

Recall, this strategy consists of socially distancing only those who fit in the category of being particularly vulnerable to the virus, and quarantining those with the infection and those living in the same households. If that were done, it would cause some economic dislocation — e.g., for the travel and tourism industries — but it would allow most everyone else to get on with life as normal, albeit while practicing good hygiene. That is the economic and social advantage of mitigation.

The medical advantage is that it leads to a rapid spread of the virus and to herd immunity, “leading to an eventual rapid decline in case numbers and transmission dropping to low levels.” Unfortunately, in the meantime, under assumptions about its transmission to vulnerable groups (because of the likely degree of contact despite encouragement of social distancing) and the number of available critical-care hospital beds, it overwhelms health services and causes many deaths.

If we could solve the medical flaw in this strategy — and that might be possible: read on — it would still face a more obstructive flaw. Governments have already committed themselves and their prestige to a bold (if mistaken) policy and invested immense amounts of political capital in it. It’s hard enough to change their minds before they’ve made such a commitment; it’s nigh impossible to do the same when they’ve bet the house on a single number in roulette. Okay, events will force a retreat to mitigation or something like it eventually. But it would require a bolt from the blue to get them to change now.

Amazingly enough, two bolts have suddenly appeared from the blue.

The lesser bolt is that, as we noticed earlier, researchers have only lately begun to point out that the Italian statistics may greatly exaggerate those deaths caused by the virus: They amount to only 12 per cent of the total number of those who died with the virus. Most died, in effect, from other causes. And that smaller death rate from COVID-19 is likely to shrink farther as the pandemic runs its course. These doubts about the Italian statistics are important because governments and the media have been treating Italy’s experience with COVID-19 as a guide to what their own countries are likely to suffer after a time lag. What if it isn’t? This question has particular significance to the U.K. The IC scientists chose suppression over mitigation in their urgent advice to the British government because they were alarmed by data they had just received from Italy. Did that data exaggerate the Italian death rates? Or did it take into account the growing doubts about them? Probably the latter, though the U.K. media have begun to follow this story only in recent days..

Even if the Italian data showed no bias, however, a third factor must be taken into account: namely, the annual death rate in the U.K. In 2018, one full year before COVID-19 was heard of, 541,000 people died in England and Wales, most of them older and less healthy people. That’s almost the exact prediction in the IC report of how many people would die if nothing was done. Are the 510,000 deaths in addition to the annual total? Apparently not. They will be part of the total. Naturally, no one now knows how large a part, since the deaths are hypothetical and the deceased still alive. But since those Brits who died in earlier years are similar in relevant respects to the great majority of Italians who died with the coronavirus rather than by it, it’s reasonable to argue that the deaths from the virus in the U.K. will not add all that many to the annual total of the dead of previous years, since many of them would likely die if the virus hadn’t erupted among us.

That’s not to dismiss the fates of human beings with a claim on us, merely to refine what is at stake. If the main aim of policy is to prevent those deaths from occurring all at the same time and overwhelming the health system, then — yes — it makes sense to adopt suppression. But if the main aim is to save their lives while avoiding an economic dislocation that would put many more lives at risk too, then that might be better accomplished by Smith’s policy of weaponized mitigation. His policy would combine paying the elderly to self-quarantine for a period while organizing industry and the voluntary services to equip hospitals with more beds and better medical technology in real time. And the latter is already happening throughout the English-speaking world.

That approach would work more easily and surely, however, if “herd immunity” were more advanced in Britain, so that fewer people would be at risk of catching the illness and therefore fearful about it. That possibility has just been delivered by the second bolt from the blue, hurled, oddly enough, by the “Pink’ un.” The Financial Times has just broken a story that Oxford medical researchers have developed a model that shows among other things that Britain has already developed a high degree of “herd immunity.”  Work by Oxford’s Evolutionary Ecology of Infectious Disease group suggests that the coronavirus could have arrived in Britain in mid January, far earlier than previously believed, spread widely under the radar for more than a month, and by now infected up to half of the population, most without their ever realizing they had been infected. Sunetra Gupta, the leader of the study and a professor of theoretical epidemiology, told the FT that, if the results are confirmed, they mean that “the vast majority develop very mild symptoms or none at all.”

If Professor Gupta’s work survives testing and, presumably, some pushback by the scientists at Imperial College, it will be a great and welcome achievement. It would mean that far fewer people are now at risk of a painful illness and death, that the balance of advantage between mitigation and suppression has now changed decisively in favor of the former, and that the destructive policy of closing down the economy to fight the coronavirus at recurring intervals can now be reversed or at the very least put on hold. As Robert VerBruggen points out on The Corner, “this is a possibility the paper sketches out, not an actual finding inferred from the data.” But if that possibility turns out to be true, it would change the entire gloomy landscape we have all felt trapped inside.

Millions of people already feel luckier. Among them, Lucky Boris, who’s been handed an alternative to trudging into a socialist prison (if he’s prepared to take it), and Lucky Trump, who’s been given the justification he needs to close down the shutdown.

And here beside the Danube, we’re no longer waiting for the coronavirus in a bittersweet mood of Central European pessimism. We’re waiting for Professor Gupta’s test results like cockeyed optimists.

Unfortunately, it just started snowing again.

Original article here.