A CSH overview on the COVID-19 pandemic


While India is slowly alleviating its stringent lockdown, CSH is thinking of slowly reopening its premises, as per the instruction of the French Embassy. While the Center has been physically closed since March 16th 2020, all our researchers are active pursuing their research agenda, some of which concerning directly the impressive epidemic that has now hit 213 countries and territories all over the World, and has killed almost 375k people. The website https://www.worldometers.info/coronavirus/ gives regularly updated statistics on the evolution of the epidemic worldwide, while a very nice geographic tracker of the quickly evolving Indian situation is provided at https://covindia.com.

Country number of cases number of deaths daily % of increase Country population (in M) deaths per million people Shut down  severity (Google mobility retail and recreation) number of tests per million people Median Age
Belgium 58 517 9 486 0.2 11.54 822.06 severe 17 147 41.4
Spain 286 718 27 127 0.0 46.74 580.42 very severe 19 905 42.7
UK 276 332 39 045 1.44 67.53 578.19 severe 9 487 40.5
Italy 233 197 33 475 0.18 60.55 552.85 very severe 28 293 45.5
France 189 220 28 833 0.11 65.13 442.7 very severe 6 823 41.4
Sweden 37 814 4 403 0.18 10.04 438.7 nothing 9 150 41.2
Netherlands 46 545 5 962 0.1 17.1 348.71 so-so 10 801 42.6
US 1 859 323 106 925 0.69 329.06 324.94 so-so 15 811 38.1
Canada 91 705 7 326 0.42 37.41 195.82 severe 17 419 42.2
Peru 170 039 4 634 2.84 32.51 142.54 very severe 6 647 28.0
Brazil 529 405 30 046 2.09 211.05 142.36 so-so 2 496 32.6
Germany 183 765 8 618 0.15 83.52 103.19 so-so 24 927 47.1
Iran 154 445 7 878 1.04 82.91 95.01 N.A. 4 930 30.3
Mexico 93 435 10 167 2.39 127.58 79.69 so-so 492 28.3
Turkey 164 769 4 563 0.51 83.43 54.69 so-so 10 445 30.9
Russia 414 818 4 855 3.45 145.87 33.28 N.A. 18 546 39.6
Poland 24 165 1 074 0.94 37.89 28.35 so-so 7 268 40.7
Saudi Arabia 87 142 525 4.37 34.27 15.32 severe N.A. 27.5
South Africa 34 357 705 3.22 58.56 12.04 severe 2 739 27.1
Maldives 1 829 6 20.0 0.53 11.3 N. A. N.A. 28.2
Egypt 26 384 1 005 4.8 100.39 10.01 severe 550 23.9
Philippines 18 638 960 0.31 108.12 8.88 severe 805 23.5
Pakistan 76 398 1 621 5.06 216.57 7.49 severe 517 23.8
Japan 16 884 892 0.11 126.86 7.03 mild 1 169 47.3
Indonesia 26 940 1 641 1.74 270.63 6.06 so-so 196 30.2
South Korea 11 541 272 0.37 51.23 5.31 mild 11 510 41.8
New Zealand 1 504 22 0.0 4.78 4.6 very severe 23 132 37.9
Singapore 35 292 24 4.35 5.8 4.13 severe 21 350 34.6
Bangladesh 49 534 672 3.38 163.05 4.12 severe 260 26.7
India 198 706 5 608 3.7 1366.42 4.1 very severe 462 28.1
Australia 7 221 103 0.0 25.2 4.09 so-so 18 993 38.7
Malaysia 7 857 115 0.0 31.95 3.6 severe 3 714 28.5
China 83 022 4 634 0.0 1433.78 3.23 N.A. N.A. 37.4
Thailand 3 083 58 1.75 69.04 0.84 so-so 726 37.7
Hong Kong 1 088 4 0.0 7.44 0.54 mild N.A. 44.4
Sri Lanka 1 643 11 0.0 21.32 0.52 very severe N.A. 32.8
Taiwan 443 7 0.0 23.77 0.29 nothing 2 561 40.7
Nepal 1 811 8 0.0 28.61 0.28 severe 1 756 24.1
Myanmar 228 6 0.0 54.05 0.11 severe N.A. 28.2
Vietnam 328 0 0.0 96.46 0.0 so-so 2 149 30.5
Cambodia 125 0 0 16.49 0 so-so N.A. 25.3
Bhutan 47 0 0 0.76 0 N.A. N.A. 27.6
Laos 19 0 0 7.17 0 severe 10 23.0

The table above, updated every day, provides numbers describing the current situation of the COVID in East Asian countries as well as in a few others that are “salient” in terms of their population size, severity of the COVID, or their geographical localization. Countries in this table are ranked decreasingly in terms of the lethality rate of the COVID-19. As can be seen, countries are extremely diverse on this matter. In the world’s most severely affected country so far, Belgium, almost 820 persons per million have died from the COVID. This number is quite impressive, even though it may be partly the result of the particular way by which the Belgian authorities are counting their COVID deaths (in which they include deaths from people who have not actually been tested for the COVID). By comparison, the total number of Belgians who die in a year from road accidents is 67 per million. At the other extreme, one finds countries like Vietnam, Laos or Cambodia who have not yet recorded any single death from COVID. The overall very low lethality of COVID in the Indochina Peninsula (including Myanmar and Thailand) is extremely surprising, and may be suggestive of a genetic characteristic, or crossed immunity phenomenon in the populations of those countries. Even countries and territories like Japan, South Korea, Hong Kong and Taiwan, who have been affected by the epidemic for quite a while now, have paid a very low death toll to this epidemic (seven deaths per million in Japan, the most affected of these countries). Our colleagues from the Maison Franco-Japonaise in Tokyo are providing regularly updated data on the progression of the epidemic in Japan at: http://covid19-ifrjmfj-tokyo1.e-monsite.com/

Countries of these tables are not in the same stage of epidemic progression, as indicated by the daily increase in the total number of death. In countries who are in the end phase of their epidemic like China, or South Korea, the daily rate of increase is extremely low (it has been zero for China in the last two weeks). On the other hand, the daily rate of increase of death is well around 5% in countries like Pakistan and Egypt, who appear to be still in a growing phase of their epidemic.  Maldives, who registered its 6th COVID death even reaches 20%. Observe that India is racing with Bangladesh in this daily increase, and the two countries have passed China in terms of COVID mortality. They are now at the verge of passing ahead Australia and New Zealand.  A regularly updated ranking of 132 countries in the world in terms of their COVID-19 mortality rate is provided at: https://www.statista.com/statistics/1104709/coronavirus-deaths-worldwide-per-million-inhabitants/

The data of this table also shows the diversity of the countries in terms of their response to the epidemic. Some countries have adopted extremely stringent measures to reduce people’ mobility (e.g. complete lockdown, school and public place closing, etc.) with the obvious aim of reducing contagion. Others have been much less restrictive in their impediments to mobility. The table provides a categorization of the severity of the observed restrictions to mobility in the different countries based on the observed mobility of people by Google (as per the website: https://www.google.com/covid19/mobility/). Very severely restricted countries are those who have experienced a visible (by Google) reduction of people mobility to museum, bars, restaurants, shopping centers and themes parks, etc.) of more than 80%. At the other extreme (South Korea), one finds countries who have not been restricted at all in their mobility (less than 20%). An objective of some of the research conducted at CSH is to estimate the impact of those mobility restrictions on the progression of the epidemic. This is an important objective given the cost of those mobility restrictions. In India, the lockdown has destroyed the jobs of hundred millions of poor people. In France, the lockdown has already cost to the country 6% of its GDP. Evaluating the impact of mobility restrictions on the the death toll of the COVID-19 is by no mean easy. Consider for example the following picture, showing the distribution of seventy countries in terms of their mortality rate and the severity of the restrictions imposed to their citizens.

Graph showing the relation between Median age of the countries and the death toll per million

It is difficult to fully understand the extreme diversity of the countries in terms of their COVID-19 death toll, and the fact that East Asian countries tend to be significantly less affected by this Virus than Western and Latin American ones. One often mentioned explanatory factor of COVID-19 mortality is the age structure of the population. The picture above, which maps the country’s COVID-19 mortality rate against the country’s median age is certainly suggestive of such a relationship, which seems to have an exponential structure, with mortality rates surging when reaching a median age of 40. While the trend seems clear, it leaves many national situations to be explained. Japan for instance, has a median age higher than Italy, and has a much lower COVID-19 mortality rate than this country. The Latin American countries are though making this relationship a bit more blurry, with relatively young populations and a growing death toll in countries like Brazil, Peru or Ecuador. But the relative protection that a low median age seems to provide against the probability of dying from the COVID-19 is a good news for country like India, whose median age is 28.5.

Countries by mortality rate and severity of mobility restrictions (according to Google, hence China and Iran are not present)

A naïve look that the picture suggests that the most severely infected countries tend to be those who have adopted the most severe restrictions to their citizens’ mobility. But the direction of the causality is far from clear. It is plausible that the most severely infected countries (such as Spain, Italy and France) have been much prone than others to adopt severe restrictions to their citizens’ mobility. It is therefore important to do a careful statistical analysis that will focus on countries having exhibited a similar initial progression of their epidemics, but who have differed markedly in terms of the restriction of their citizens mobility that they have decided to implement. Good examples of such countries are Belgium and Netherlands, or Sweden and Denmark.

Other researchers at CSH, such as Olivier Telle and Samuel Benkimoun, are also examining the impact of the countries restriction to mobility using Facebook data, available at an extremely precise geographical scale. They plan to connect these local restrictions to mobility to the local spread of the disease so as to obtain a more precise picture of the impact of the mobility restrictions on the spread of the virus. Some other are eager to understand from a more qualitative perspective the impact of the Indian lockdown on the underprivileged segments of the Indian population.

The Indian situation:

India has now been under a tight lockdown for more than seven weeks, even though it has been significantly alleviated since May 3rd. While this lockdown has hurt heavily the Indian economy and, in particular, the fragile segments of its population, its effect on the epidemic dynamic is far from clear. The following picture shows the cumulative number of COVID deaths in India since the beginning of March. As can be seen, this cumulative number has been growing at a slightly increasing rate since the beginning of the lockdown. There is until now no evidence that this growth is slowing, and it can even be observed that the the growth has been slightly increasing 21 days after the beginning of the lockdown.

Death due to COVID-19 in India since lockdown implementation

This is in sharp contrast to the French situation, depicted on the following picture. While cumulated COVID deaths in France have been by far larger than those observed in India, the lockdown imposed by the French authorities on March 14th have clearly reduced the growth rate of those deaths 21 days after its implementation. The fact that no such effect has been visible so far for India may make one doubtful about the actual effect of the Indian lockdown on COVID mortality. 

Death due to COVID-19 in France since the lockdown implementation

While cumulated reported COVID deaths in France have been by far larger than those reported in India, the lockdown imposed by the French authorities on March 14th seems to have reduced the growth rate of those deaths 21 days after its implementation. The fact that no such effect has been visible so far for India may make one doubtful about the actual effect of the Indian lockdown on COVID mortality. 

But the measurement of COVID mortality is, in India just as elsewhere, dependant upon the number of tests realized. After all, a reported death from COVID is nothing else than a death reported in a COVID-positive tested person. As shown in the picture below, it is undisputable that India’s testing capacity, which was very low in March, has increased tremendously since the beginning of the epidemic.

It is therefore possible that the increase in the number of COVID deaths in India observed in the last three months be largely the result of the increase in the number of tests performed, rather than the increase of the spread of the COVID per se. In the following picture, we depict the estimated number of COVID cases in the Indian population targeted by the tests, based on the following formula:

# of reported COVID cases at day t = (Number of tests done up to t)

                                                                       x (fraction of the targeted population that is infected)

Since we know both the number of reported cases at day t and the number of tests performed up to t, we obtain the fraction of the Indian population targeted by these tests that is contaminated as the ratio of the two. The graph clearly shows a stabilization of this fraction after the lockdown, even though an mild upsurge has been observed in the last 10 days. Overall, these estimates suggests that roughly 5% of the Indian population could have been contaminated by the COVID so far.

Nicolas Gravel, Director of the CSH.

You may also like...