Thursday 17 September 2020

COVID-19: Is the Virus Weakening?


This is another of my articles on the numbers relating to the COVID epidemic world-wide.  It follows on from the “Lock-downs or Cock-ups?” article, here: https://wattsupwiththat.com/2020/08/11/covid-19-lock-downs-or-cock-ups/. This time, I’ll focus on the question: what proportion of confirmed new cases, at each stage of the epidemic, are actually leading to deaths? That ratio ought to be a major factor in any rational consideration of when to release lockdowns (or not); because it hugely affects the load on health care systems. Avoiding health system overload, so we were assured at the beginning of the epidemic, was the only reason for going into lockdown – for three weeks. (Cough).

Again, my data comes from Our World in Data, https://ourworldindata.org/coronavirus-source-data. The data for this paper was taken on, and runs up to, September 12th. Once again, the stringency data came from the Blavatnik School of Government, https://www.bsg.ox.ac.uk/research/research-projects/coronavirus-government-response-tracker. The detail file I used was the version from September 13th. I’ve commented earlier about the shortcomings of this data; but as I’ve said before, it’s the best data I’ve got, so I’ll use it.

My plan of campaign

I thought I would organize this article the same way as my first major article on the subject, back in June. That is, looking in turn at different regions of the world, with a section on the world-wide situation as well. I couldn’t afford the time to tackle all 213 affected countries (nor could you afford the time to read the result!) So, I did the following:

1.     I started with the 15 countries with the highest death rates per million, as of September 12th. That gave me: San Marino, Peru, Belgium, Andorra, Spain, Chile, Bolivia, Brazil, UK, Ecuador, USA, Italy, Sweden, Mexico, Panama.

2.     I added those, among the 12 countries with the highest number of new cases on that same day, which weren’t already in my list. That added: India, Argentina, France, Colombia, Russia, Iraq, Philippines.

3.     I added those, among the 10 countries with the highest cases per million rate, which I didn’t already have. That gave: Qatar, Bahrain, Aruba, Kuwait, Oman.

4.     To ensure I had a world-wide mix, I added eight countries which received the virus back in January or February, and which weren’t already on my list. They were: China, Thailand, Japan, Taiwan, Australia, South Korea, Vietnam, Iran.

5.     I added a number of countries on which I had remarked in earlier articles. That gave me a total of 52 countries – around a quarter of all those affected – plus an entry for “World.”

Oh, and I looked at the Faeroes as well. But, beyond the fact that they have got on top of their second wave, though rather more slowly than they got on top of the first, there’s nothing to report from there.

The main decision I had to make was how big a lag time to allow between a new confirmed case and the death which might or might not result from that case. I ran across a paper from the Centre for Evidence Based Medicine, which suggested that the mean time between confirmed case and death in the UK is of the order of 17.5 days; but also, that you should be able to pick out trends using a lag time of 14 days. So, I decided to keep it simple, and compare the numbers of COVID deaths in each country with the number of new confirmed cases reported 14 days previously. Of course, that lag time applies to the UK; it might be different elsewhere.

And because the raw data is very noisy, I weekly averaged both deaths and cases before dividing the one by the other.

The UK

Just recently, we had some news in the UK that wasn’t fake. COVID cases are going up!

Now, the ruling classes – in the UK, as everywhere else – only tell the truth when it suits their agenda. Which, in this case, is yet more tightening of the lockdown restrictions; in this case, restricting the maximum allowed group size of a gathering to six. Maybe they did this out of spite, because of the demonstration that Piers Corbyn (older brother of Jeremy, until recently leader of the Labour opposition) organized in Trafalgar Square a couple of weeks ago against the lockdowns. He’s even being called a “coronavirus denier!”

And what is a “case” is, it seems, a movable feast. Put out test kits with a higher false positive rate than earlier ones; et voilà! You can have as many new “cases” as you want.

But there’s a heart-warming graph to come next:

There’s a surge in late March and April, up to a plateau between 20% and 25%. The ratio of deaths to new cases from 14 days earlier peaked at 26.4% on June 10th, before starting a steep descent. As of September 9th, it had fallen all the way to 0.4%. The virus’s lethality in the UK has gone down by a factor of 60 over the course of just three months! That’s really good news.

Moreover, “The Week” (https://www.theweek.co.uk/107790/flu-killing-more-people-than-covid-19) reported a month ago that, by the end of July, ’flu was killing five times as many people in the UK than COVID was. Suggesting that there is now no objective reason for any lockdown at all.

So, why have the deaths per case figures fallen by such a big margin? Is the virus weakening? Has it taken most of the “low hanging fruit,” so those who are getting it now are healthier, so less likely to die from it, than earlier? Or could it be, perhaps, that testing has expanded by orders of magnitude over the last three months? Or that the latest generation of test kits has a higher false positive rate than the ones which were used at the beginning of the epidemic?

I can deal with two of these possibilities. The number of tests recorded from September 1st-10th inclusive was 1.82 million. Against the number from June 1st-10th inclusive, 0.67 million. The ratio is just over 2.7; significant, but nowhere near 60+. And even if new kits record twice as many positives in a given population as the older ones did (and that would be a serious violation of any half decent set of quality standards), that still doesn’t get us even within a factor of 10 of the improvement we see. So, the drop in deaths per case is real, not an artefact of the measurement process. Thus, it ought to be taken into account in policy decisions.

Furthermore, as of September 17th, I see no clear evidence of any increase in deaths 14 or so days after the surge in cases in late August and early September.

So, why tighten the lockdown? They should be opening up, not locking down harder. Many people have gone without normal human social relations for six months now, and people are getting tired, stressed and angry – not to mention short of money. But that’s typical of the political class – never let a good crisis go to waste, particularly when it’s only the “little people” who are suffering. Meanwhile, bureaucrats and “key workers” are being paid overtime!

Now, they have recently re-opened the schools, and that may be the main reason behind the recent surge in cases. If so, isn’t that a good thing? As long as the hospitals aren’t overloaded, won’t going for herd immunity – which we were told was the original strategy, weren’t we? – get rid of the damned thing with minimum pain from here on in?

As to compulsory “face coverings,” they are counter-productive. Nothing less than a professional-level mask will protect the wearer. And if the idea is to protect everyone else, then you’d do better to take the things off and let the virus run its course towards herd immunity.

I think the main difficulty is that the establishment, and the bureaucrats in particular, enjoy their power to make people kow-tow. They will tend to advise for restrictions rather than against. And they will hold out against any opening up, even if the politicians wanted to do that. As I know from having worked with a few of them, government managers tend to be risk averse. They won’t take a risk, if it might make them look bad. But the problem is, the awful political system we suffer under doesn’t put them under any pressure in the opposite direction. If what they do (or fail to do) causes real problems to us “little people” out here in the real world, they don’t get held accountable. The worst that can happen is that their political masters will be booted out at the next election; but they themselves will still have well-paid jobs, cushy pensions and power.

This lack of accountability tends to result in bad situations being perpetuated, for far longer than is reasonable. A good example was food rationing in the UK after the second world war. It lasted for eight whole years after VE Day – longer than the war itself! If Johnson or someone doesn’t start bashing some heads together, I can foresee us still being under lockdown in two years’ time.

But it’s worse than that. There is a significant faction within the UK ruling class, that wants to use the COVID virus as an excuse to impose their own destructive agendas on people. This is the same faction that, last year, kow-towed to Extinction Rebellion, and instituted “citizens’ climate assemblies.” They hate us “little people,” and they hate our Western industrial civilization. Unsurprisingly, they want the lockdowns continued as long as possible, even intensified. In their Utopian dreams, they’d like to make the restrictions permanent. And they have made regulations that let them severely punish anyone who objects – like Piers Corbyn. As shown by the £10,000 fine they slapped on him after his London protest.

Northern Europe

Next, I collected together the deaths and deaths versus cases data for a group of eight Northern European countries. Here are the cumulative deaths per million.

As you can see, things have been very quiet indeed for several weeks. New deaths from COVID are minimal in comparison to a few months ago. And with the possible exceptions of Belgium and Luxembourg, they all look ready to call off the lockdowns, and get back to living again.

Now, here are the ratios of deaths to cases (with a lag of 14 days) for those same eight countries. Spaghetti warning!


The French have given us a passable imitation of the façade of Notre-Dame cathedral; but then, I haven’t trusted any of their figures, right from the start. The Belgian peak was horribly high. Why the Dutch peak was later than the others, I don’t know. Maybe the virus got into the care homes later there than elsewhere? The UK and Sweden – and, interestingly, Germany – have bumbled along somewhere in the middle. The Icelanders are thumbing their noses at the rest of us. And Luxembourg… some strange things happening there recently. Maybe the praise I gave them a month ago went to their heads.

Let’s take a closer look at the last three months.


So, the UK is not alone in having got on top of the virus. All eight countries are at or below 1% deaths per case now.

There is only one of these countries for which I feel a need to follow up: Luxembourg.


Despite the unsightly adjustment, I see nothing to worry about there. And yet, after I applauded them a month ago for their excellent lockdown release strategy, they have started to impose the lockdowns again! I wonder why? Political pressure, perhaps, from neighbouring countries?

Southern and Central Europe

Here is the data from seven Southern and Central European countries:


Spain and Portugal are still moving upwards a bit; the others all look good.


The Spanish data is marred by two large adjustments. But otherwise, apart from Andorra’s nightmare, we see many of the same features here as in Northern Europe.

Let’s look at Austria first.




The Austrians seem to be managing things as well as anyone. Lots of new cases in September, coinciding with back-to-school time, just as in the UK (and in the Netherlands too). But also, fast-dropping deaths per case since the second peak in June, and a commendable reluctance (so far, at least) to panic and re-impose lockdowns once lifted. And they have (as we’ll see later) the fourth lightest lockdown in the world as at the end of August!

The Italian new cases per day has a similar shape to the UK’s, and the deaths per case has a similar shape too, but shows a far later peak. In Portugal, the new cases graph has followed a similar path to other European countries, but the deaths per case, while never going as high as in Italy or Switzerland, has been relatively slow to come down.

Spain is the first country I’ve seen to have gone above the daily new case levels from the first peak.


Unfortunately, due to the adjustments you can see on the deaths line, the detailed deaths per case data for Spain is meaningless. But as the comparison with the other countries shows, their deaths per case now (September) is less than 1%. And yet, the political trend in Spain since late June has been towards locking down again.      


I wonder if they are taking too much notice of the new case count, and not enough of the (lack of) deaths?

To sum up where the Southern Europeans are, here are the deaths versus cases for the last three months, excluding Spain.


As of September 9th, all these countries had deaths per case ratios below 1%; and neither Andorra nor San Marino had had a COVID death for some months. So Western Europe as a whole appears to be all but recovered.

The world

Time to take a step back, and look at some world-wide data. The individual graphs of cases and deaths are easily available elsewhere, so I’ll just show you the deaths per case:


You can clearly see the Chinese peak in February, and the European one in April. Since then, it’s been downhill; but it seems to have levelled off since July.

Meanwhile, the “league leaders” in deaths per million population, cases per million population, and deaths per case, are as follows:




All this data is as of September 12th. Now, Yemen is an outlier, being a war zone. But otherwise, deaths per case is a good measure of the lack of quality in a country’s health care system. For both insufficient testing resources early in the epidemic, and poor standards or the wrong treatments in hospitals, will drive up the deaths per case.

Italy, the UK and Belgium all have health care systems over which government keeps close control. Germany, on the other hand, is known for its consumer-oriented health care system. It would be interesting to compare the performance of each of the European health care systems over the COVID outbreak, against the level of political control!

Now, to lockdowns. Here are the harshest locked down and least locked down countries in the world, on the Blavatnik measure, as of the end of August.



Austria, Iceland and Luxembourg are notable contenders for the best handling of lockdown release. Our Swedish friends, I think, have made very clearly their point that stringent lockdowns were a bad idea in the first place. And San Marino and Andorra, obviously, have had few recent cases, since the virus is already all but “burned out” there. Taiwan, though, is surprising; particularly given that China across the water is the eighth hardest locked down country in the world!

And here are the horror stories – the countries with the highest lockdowns at their peak.


Ouch!

North America

Across the Atlantic now. First, let’s have a look at the deaths per million and deaths per case for the five countries from my sample which are in North (and Central) America.


Canada looks OK. Panama has just about started on a downturn in death rates. But the USA, Mexico and Aruba are all on the way up, with no bend in sight.


Canada looks to have a similar profile to the European countries. The USA started out heading that way, but suffered an upturn again in July. In Mexico, deaths per case are high, and there’s no sign yet of them coming down; ouch! And Panama seems now to have settled on a similar trajectory to the USA.

Now all of a sudden, a country I hadn’t looked at before, Aruba (a Dutch dependency in the Caribbean) has rushed into third place in the cases per million stakes. In the graph above, you can just see it as a little blue spike in the bottom right-hand corner. So, what’s going on there?



So, they had a relatively easy first wave, thought it was all over, started unlocking, then got hit by a second wave at the beginning of August. This may, perhaps, be evidence that the virus genuinely has weakened over the months. If Aruba had got this thing early – as Sint Maarten did – would they have got off so lightly? Today, Sint Maarten’s deaths per million is more than twice Aruba’s.

You can see how badly “in the wars” Mexico is from the daily cases and deaths below, and the grey line in the deaths versus cases graph above.



As I remarked last time, nothing the Mexicans have tried to do to control the virus seems to be having any effect – either way. The organism is just doing as it damn well pleases.

Panama’s daily cases and deaths graph looks much like Mexico’s, but it has consistently lower deaths per case. Both countries have a combination of state and private health care, but the Mexican state system (which is a similar age to the UK NHS) has a bit of a chequered history.

As to the USA, deaths per case recently are significantly higher than in Europe. This may well be due to the virus still spreading across the country into previously unaffected or lightly affected states and populations. There’s obviously still a long way to go.

South America

To South America, the continent worst affected by the virus at present. Apart from Chile and Brazil, all the seven countries shown below have been subjected to lockdowns above 90%. And as at the end of August, Argentina, Bolivia, Chile, Colombia and Peru were five of the top seven heaviest locked down countries in the world. Moreover, all of them have high ratios of cases per test, suggesting they have a shortage of testing capability.

So, here are the cumulative deaths per million for the seven South American countries in my sample.


The quality of the data from South America isn’t good. You can see, above, two big upward deaths adjustments for Peru, two smaller ones for Chile, and a large one each (on the same day) for Bolivia and Ecuador. And that’s only the deaths; the case counts may not be so good, either. So, the deaths per case data may not be all that meaningful. Ecuador’s data, indeed, is so bad I had to discard it. But let’s have a look at the rest.


Among all the spaghetti, I get a sense that many of these countries are going the same way as Panama and the USA. Up to a peak, down and then back up again. Bolivia is in big trouble, like Mexico; all the others are currently running at around 2% to 4% deaths to cases. Which, to be fair, is far better than the Europeans near the peaks of their epidemics. But only Chile shows any recent downward movement.

For more detail, I’ll look at Brazil first.


At least they’re “over the hump” in terms of new cases. That’s something. And the lockdown graph again shows the “Mexico effect.” It seems to be characteristic of Latin America that even the heaviest lockdowns have had no discernible effect in either direction.


Argentina’s cases are still on the way up. But their deaths per case graph looks not dissimilar to Brazil’s (blue line above, against the grey).

All the figures from Chile and Ecuador, and the deaths from Peru, are marred by adjustments so large that they render the data all but meaningless. Peru’s case counts, on the other hand, show them having “turned the corner.” And lastly, Colombia shows a pattern similar to Brazil.

Australasia

Let’s go round the world the other way this time; to Aussie and NZ.


Both these countries set out their stalls to keep the virus out, and let other people deal with it. A rather questionable strategy, in my view. So far, New Zealand has been the more successful of the two.


Australia has had a two-phase epidemic. In contrast to Europe, the second phase looks more lethal than the first. Have they been too successful in halting the first phase quickly? And Australians remained locked down for two months or so, during which there were few cases and very few deaths. A whole lot of pain for little, or perhaps even no, gain.



New Zealand also has had a two-phase epidemic. They released the domestic lockdowns, though not the draconian border restrictions, for a couple of months. But they brought them in again, on a regional basis only, for the second phase.


South-East Asia


Deaths per million here are low by the standards of Europe or the Americas. These countries divide into three groups. Indian deaths are still rising almost exponentially. Bangladesh and Indonesia are at about the same level; but Bangladesh’s curve has straightened out, while Indonesia’s hasn’t yet. Singapore, Thailand and Vietnam have had only 27, 58 and 35 COVID deaths respectively. Which makes me wonder, is there a benefit against a virus like this to having a Chinese heritage?

Here are the deaths per case. They are rather distorted by the fact that the outbreak in Vietnam was both recent and quickly snuffed out.


To make things clearer, here are the ratios for the last 3 months, with Vietnam removed.


In India, new cases per day are still rising. But the deaths to cases ratio is now down to not far above 1%. That’s good news.

Indonesia is on a similar trend to India, but with a lower number of cases per million and a somewhat higher deaths per case. Bangladesh has “turned the corner” in number of cases; but its deaths per case ratio has been, very slowly, edging upwards since May.

Singapore cases have gone up and down like a yo-yo, though they are now down into double figures per day. But they have had so few deaths (only 27 in the whole epidemic) that the deaths per case figure is pretty much zero. Thailand had basically beaten the virus by the end of April, and they haven’t had a death since June 3rd; job done.

But Vietnam is more interesting. They have had a multi-phase epidemic, without any deaths until August. The apparent runaway in deaths per case in August and September is simply due to the small number of cases. The reality shows better in the cumulative deaths per case graph, below.




East Asia

Under the heading “East Asia” I have put: China, Japan, the Philippines, Russia, South Korea and Taiwan. Including Russia here is perhaps a little unfair; but it’s a big country, and it does have a substantial border with China.

To China first.




That looks to me a high lockdown level to be maintaining for such a small number of cases! When I looked at the Blavatnik data, I found that the national lockdowns ended way back in March; so apart from a ban on international arrivals, these lockdowns are all at the provincial level. Not that we can believe the Chinese on coronavirus matters, of course.

But that raises a question mark against the Blavatnik methodology. If one province out of many is locked down, how much should that count towards a national lockdown index? “According to its population” is the obvious answer. But that isn’t what they’ve done.

Now let’s look at the other five together.


As I expected, Russia is an outlier here; its deaths per million now being comparable to countries like Germany. And it’s still going up, though it does look to be on the second half of the S-shape. The death toll in the Philippines is still rising, but from a low base. The others all have low deaths per million.

Let’s look at Japan.



The lag from cases to deaths in Japan seems to be quite a bit longer than 14 days, which explains the peak being in May, when cases were right down. Here, also, there is an epidemic of two phases; the first more lethal than the second. As of September 9th, the deaths to cases ratio was 2.3%.

The Philippines, too, have had an initial lethal outbreak, followed by a much larger and far less lethal one.

South Korea looks comparable to Japan; a lethal outbreak in February, followed by a smaller and far less lethal one in August.

And Taiwan? Less than 500 cases, and only 7 deaths, in the entire epidemic so far. And they have, on the Blavatnik measure, the least stringent lockdown regime in the world! If you’re Taiwanese, that is; for the borders are all but closed, and are only being opened up slowly to business and student travellers, with mandatory quarantine. I also wonder, does their Chinese ancestry help them against a type of disease that has, very probably, been circulating in that part of the world for centuries?

The Middle East


This doesn’t look very good. Iran is in trouble, with no sign of any slackening in the death toll. Iraq, too, has deaths per million higher than countries like Portugal. At the other extreme, Pakistan seems to have “got away with it.” Most of the rest are in a linear phase, with deaths rising at a roughly constant rate.

How did Pakistan achieve this? The best information I could find is that the virus has been confined almost entirely to the biggest cities. The Blavatnik data tells me that there have been mandatory travel restrictions for much of the time, and people are still being encouraged to stay at home. So, the actual effect of the lockdown is a lot more stringent than the 48% figure suggests. I suspect that, when they finally open up internal travel, they may find their initial success coming back to bite them.

This puts another question mark against the Blavatnik methodology. Forced stay-at-home is well more than 80% lockdown in my opinion, even if not accompanied by any other measures at all.

The deaths per case spaghetti, on the other hand, is all over the place.


Iran and Iraq have both had two-phase epidemics, which Iraq seems to have got through more effectively. Saudi Arabia’s deaths per case are still climbing.  Oman got the virus relatively late, and has now reached a deaths per case of about 2%; comparable with Pakistan, Iraq and Turkey.

The relative success stories are Bahrain, Kuwait and Qatar, all of which have managed to reduce their deaths per case rates below 1%. They are all in the top 10 in the world in cases per million, so I suspect this is down to a strong testing program.

South Africa

I’ll conclude with a geographical outlier; the only country in Africa which qualified for my list.




It’s bad at the moment, because they’re at the top of the “lethality hump.” But I think they’ll make it.

Conclusions

Western European countries, in my view, are all but recovered from the virus. With enough cases per million, and so enough immunity, to make a highly lethal “second wave” very unlikely. We must fully re-open international travel and the economy in Europe as quickly as possible. And, of course, we must get rid – permanently – of all the bad regulations that were made in the name of fighting the virus.

What I would like to see is the different countries in Europe co-operating against the virus. Each should try a different combination of unlocks; monitoring the results, and seeing what works and what doesn’t. Then what works in Luxembourg, say, can be tried in France. What works in Switzerland can be tried in Germany. And what works (and has worked all along) in Sweden can be tried in the UK. This should return Europeans to some semblance of normality as fast as possible. Then, we can start on the long road towards economic re-building. Moreover, such a friendly co-operation would show up for all to see the superfluity and worthlessness of the EU.

Outside Europe, only Canada, Bahrain, Kuwait and Qatar have the deaths per case rate down below 1%. People in most of the rest of the world now find the virus spreading around them, but nowhere near as lethally as it did in Europe. A 2-3% case fatality rate seems to be the norm. That’s bad; but nothing like what happened in Belgium, the Netherlands or the UK. People in these places should simply be able to get on with their lives. There may be some countries – such as Mexico, Bolivia, Iran, perhaps Saudi Arabia – which will need to continue harsh restrictions for a time. But there should be no more than a few of these. And the countries that have shut down travel – Taiwan, New Zealand, Pakistan, to name but three – need to open up again.

2 comments:

Opher Goodwin said...

Exceedingly thorough Neil. I'm a little sceptical about the quality of a lot of the data though. I don't know how realistic a picture it paints.
As to the death rate and the second wave - I think we will have a big surge in deaths again. Time will tell.
Interesting stuff.

Neil said...

Opher, I'm even more skeptical than you are about the quality of the data. All of it comes, at one level or another, from governments. And governments are not trustworthy.

Time will, indeed, tell.