Since the death rates due to COVID 19 seem to vary strongly between different countries, the suspicion stands to reason that this may have to do with different qualities of the health care systems in these countries. So I have compiled a list of countries ranking them based on the ratio of deaths due to COVID 19 and cumulative infected persons up to 10/04/2010 (around noon) according to data published by the Johns Hopkins University. To make sure that these are countries not just at the beginning of an outbreak (where rates may not yet be stable), I have restricted myself to cases with at least 20000 cumulative infections, except for South Korea and Austria, where I happen to know that they have a sufficiently long history with the disease. This gives the table below, in which I have also included the number of reported recovered cases, which gives an indication as to the current phase of the pandemic in the respective country. And I have added data for the city of New York which alone has more cases than all of China (if the numbers given by China are correct).
Date: 10/04/2020
Country Cumulative Infected Deaths Rates (%) Recovered
--------------------------------------------------------------------------------------------------------------
South Korea 10450 208 1.99 7117
Turkey 42282 908 2.14 2142
Germany 118235 2607 2.20 52407
Austria 13377 319 2.38 6064
Canada 20765 510 2.46 5311
US 466299 16686 3.58 26522
Switzerland 24172 958 3.96 10600
China 82924 3340 4.03 77758
New York 87028 5150 5.92 0
Iran 66220 4110 6.21 32309
Spain 157022 15843 10.09 55668
France 118785 12228 10.29 23441
Netherlands 21910 2405 10.98 280
Belgium 26667 3019 11.32 5568
United Kingdom 65872 7993 12.13 359
Italy 143626 18279 12.72 28470
(Because this editor does not reproduce spacings correctly and the text font width is not fixed, I attach the table as a pdf document and as a png image, which will make it more readable.)
As an aside, it may be noted that the ratio of the number of deaths to the number of recoveries for the closed cases, i.e., the known cases of infections that have meanwhile ended one way or another, is much higher as of April 8. Deaths constituted 21.12 %, recoveries 78.88 % (source: https://www.worldometers.info/coronavirus/worldwide-graphs/#newly-infected-newly-recovered) of all such cases. This does not mean that in the long run such a high percentage will die, because this count misses all recoveries of persons that had too weak symptoms to ever enter a hospital or get a test.
The table contains some expected and some unexpected results. For example, while the relative positions of South Korea and Germany were roughly to be expected for someone following the development of the last few weeks, the position of Turkey is a real surprise to me. But seeing that they have only one twenty-fifth of recovered cases at about a third of infected persons in comparison with Germany, I would conclude that they are still much closer to the beginning of the epidemic than Germany and they either test a lot or have a real explosive development of infections, increasing the denominator in the ratio deaths vs cumulative infected. Their rate of deaths vs recovered persons is not particularly good so the conclusion that they have a good health care system would be premature. But we shall see whether they can keep their position...
Germany, Austria, and Canada seem to fare about equally well. The pandemic started somewhat earlier in Austria than in Germany and their infection curve seems to flatten more strongly already. But the numbers seem to correspond to the fact that Austria has about one tenth of the population of Germany.
The US have, to my knowledge, the most expensive health system in the world, about twice as expensive per capita as th German one. But it is a system mostly for the rich. If you need treatment for certain diseases where cutting-edge research results are important (such as certain cancers), getting that treatment in the US may be life-saving. If you can afford to pay for it, you'll have your surgery or whatever in the best clinic which likely is in the US. But many US citizens will not benefit from the advances at the top of their system, because they have no health insurance at all or no insurance at the moment of disease, their insurance being coupled to employment. So it may not be too surprising that the US does not show up at the top of the list. On average, they are still doing pretty well, but the social component of their system becomes immediately visible by looking at New York that has about the same quality of results as Iran... (Also, there are rumours that black people have lower survival chances in New York than white ones.) The figures for Iran may not be reliable, though.
There may be a large dark figure in the case of China, too, so it is not clear whether their relatively low rate signifies that they have health care comparable to Switzerland indeed.
In the case of Spain and Italy, we can be sure that their health systems had at least a partial breakdown (which is probably true of France as well), which lead to an increase of death rates. There is evidence of triage in Italy, e.g. the story of a 72-year old priest who renounced a respiration apparatus in favour of a younger person -- and died. One may of course see this as a consequence of insufficient resources of the health system. Of the European countries, Germany seems to have the highest number of intensive care beds and the largest reserve capacities. Apparently, they exceed those of the UK by a factor of 10 or so, and this may play an important role. In any case, the National Health Service of the UK does not have the best reputation. And the German health care system has not yet touched its limits. This may happen two weeks from now, if the rates of new infections don't drop.
Of course, there may be other reasons for the high rates of the six European countries at the end of the table. Maybe the number of unreported cases is much higher there and the actual death rate correspondingly lower. It seems that testing for the disease is not done as extensively in those countries as in Germany. But it is doubtful whether this can account for a factor of five in the rate. Also, Germany did not test as much as South Korea (and did not as efficiently track the contacts of established cases), otherwise they would never have had 10 times as many infections. And you cannot easily get a test in Germany. Symptoms such as fever or cough are not enough to be tested for COVID 19, you need to have had contact with an infected person.
I have seen dark figures of infections estimated to be as high as six (for the US, which would mean they have 3 million rather than half a million infected persons), so it is not excluded that a substantial part of the differences is due to a wrong denominator -- assuming that the number of deaths is about correct, i.e., there are only few misattributions. But if six is a decent figure for the US, then two would probably be a minimum for the European countries (they will not be that much better in detecting infections than the US), so the spread is still only about three, leaving room to attribute differences to differences in health care quality.
Does this sound reasonable or are there more important reasons for the differences? (Age structure comes to mind, but is definitely not sufficient to explain the difference between Germany and Italy.)
Kassner,
to my opinion it is a bit too early to make assessments in this matter, since not all countries are aligned with the same level of outbreak, or rather can't be compared similar situations ( the R0 is very different in different countries). It is difficult to make comparisons, but it is good to think about it.
So far one of the highest mortality rate is the one of Netherlands as you can see from John Hopkins University table, is it one of the worst NHS in Europe?
I would suggest to make an image out of the table you exported from excel and upload the image, otherwise the numerical figures are barely intelligible (same as what I've done).
For sure there is an influence on the quality of NHS:
beginning from home care and clinical care: regular, intensive care .
For example the difference is evident in Italy from region to region, but consider that the best NHS was considered the one of Lombardia now in the deepest trouble.
There are unknown factors which have non trivial influence on mortality like:
a) the actual spread of the disease for each country: how many people got in touch with the virus and they are immune now, not counted. The number reports only the ones who got analysed which is a tiny few in general compared to the whole population (Italy has a lot of these people who were not analysed), this is strictly epidemiological.
b) the number of people who complained to have symptoms but the NHS did not consider them, they got worse at home and arrived too late at the hospital (this depends on the quality of the NHS)
c) the number of people who died with the virus at home but they were not counted positive (Cina has a lot, this depends also on the response of the NHS but not only).
and the following but known have to be considered in the analysis,.
a) the number of people hospitalised for each country, being positive to the virus
to have an idea also of how many spent their time at home not resorting to Clinical care.
b) the number of people in intensive care for each country, being positive to the virus in comparison to the maximum capacity of the structures.
Last but not least the average age of the people who died in comparison to the average age of people in the country:
this tells you to which extent the weakest have been touched by the virus in the country.
Stefano Quattrini
I have added a png image of my table. But I don't know how to put it in the place of the table in the text, which is what I actually would like to do. I just can attach an image file just as before I attached the pdf.
As to the R0, this should not be very different in different countries. Of course, social life is not hundred percent the same in every country and I would expect a slightly higher R0 in a country such as Italy than in Germany, where people are more reserved in social contacts. But considering that the cutting-off of many social contacts that is practised now is barely sufficient to reduce R0 from 3 to 1.1 (and hopefully below 1 in the long run), I don't think that the differences between social behaviours in Europe (before lockdown) will vary R0 by more than 0.2.
You are certainly right that it is too early for a final assessment. Also the fact that places in Italy where the health system was actually considered to be superior did not fare too well in comparison with places where it was not considered as good, does in itself not say more than that the break-down limit was reached. After that, the quality of the health system does not count anymore. People are dying because they do not get health care at all, not because they get worse health care...
Of course, inside a general assessment, the capacity of a health s ystem will also be a point counting towards its quality. In Germany, the pandemic is probably roughly two weeks behind that of Italy, which means that we could take slow-down measures earlier. Also, we were able to double the number of vacant intensive care beds within the last few weeks. Since even before that our number was higher per capita than elsewhere in Europe, there is some hope that there will be no explicit break-down of the system...
Anyway, here is another question regarding COVID 19 that you may find interesting:
https://www.researchgate.net/post/How_to_continue_coping_with_COVID_19
Strangely, no answer yet...
Dear K. Kassner
>
unfortunately this RG system does not offer us such feature....
Germany is doing an order of magnitude more analyses of covid-19 than us. For this reason you are probably close to the right number of people who are infected. In Italy it is very likely that the number of infected is 4 times the one analysed, that would also justify the high death rate. Also this is a consequence of a weaker NHS .
>
I'm an quite confident that Germany will be far from any breakdown of NHS.
Stefano Quattrini
"Germany is doing an order of magnitude more analyses of covid-19 than us. For this reason you are probably close to the right number of people who are infected."
I suspect that we may still be a factor of 2 low. For the US I have seen a suspected dark figure of about 6.
You do not get a test here, if you just have symptoms that look similar to those of COVID 19. You will be tested, if in addition you probably had contact to someone who was infected. My wife, who is a medical doctor, was sent home from the hospital where she works when she had a slightly raised temperature (still below the level where it is considered fever). When she asked to be tested at the local health authority, she was told that normally one would not test her with only these symptoms. Only after she pointed out that she would not be allowed to continue working in her medical profession (which would reduce the work force against the virus) without being proven virus negative, was she admitted to the test. And of course, she was not infected, so she may resume work now.
Hello K. Kassner and Stefano Quattrini ,
One of the things you do not see discussed is the FP (False Positive) and FN (False Negative) conditional probability estimates and their respective confidence intervals and confidence levels for the COVID-19 tests or the antibody tests. Without such metrics how do you really know who is and is not infected and who was infected? I have been interested in these metrics from a different point-of-view: LS NII (Large Scale Non-Intrusive Inspection) systems used to detect contraband at ports of entry by examining radiographs of cargo containers, see https://www.researchgate.net/project/Testing-Tests-2 , but these metrics are also traditionally associated with medical diagnostic tests. As many people have said: "The only thing worse than no test is a bad test." You can have Big Data, but "Big" doesn't imply better: garbage in, garbage out.
Regards,
Tom Cuff
Thomas Cuff
Hmm, I thought that these statistical problems were mostly related to diseases that are extremely rare, so that a small probability for a false-positive test would lead to (possibly dangerous) unnecessary treatments. For example, if a test is false-positive with a probability of 1/1000 and never false-negative and the disease is present with a probability of 1/10000, then out of a million tested persons you will find 1100 that seem to be ill, but only 100 are in reality. So the probability for a person getting a positive test to be really ill is 1/11. Starting a treatment with every positively tested person would be disastrous, if the side effects are severe, as you would make 10 persons suffer for each one that you try to heal and if one out of these 10 dies due to the side effects, screening tests of this kind might do more damage than good. But of course, one would test every positively tested person a second time or in a different manner, before actually starting treatment.
Now with COVID 19, we don't precisely have this problem, as the disease is no longer rare, and the consequences of positive testing is only quarantine as long as no treatment is available. Also, when symptoms appear, treatment may be necessary without prior testing.
But maybe there are other negative consequences of false positive and/or false negative tests that would have to be taken into account here? Otherwise, the problem would seem less urgent than that to get a sufficiently large number of tests running.
I suspect that outcomes from Covid-19 will prove to be a fairly weak predictor of the quality of healthcare systems overall because western healthcare systems are heavily geared towards multiple morbidity, chronic disease, elective work etc; not to mass outbreaks of infectious disease. It depends what your quality measures are of course. If quality is measured by preparedness for such an outbreak it is fairly obvious which countries would score highest but they would not perhaps score so highly on other parameters. France, Germany and the Netherlands have very comparable healthcare systems in world terms but different profiles with this disease. Also we don't yet know the long term outcomes of treatment in ICUs. These may not prove very good but let's see. Generally I agree with colleagues who have implied that "it is too early to say". And what about services' ability to cope with other illnesses during the outbreak and also with the recession afterwards. You can't look at Covid in isolation in terms of healthcare responses. I would suggest we wait. However, I would commend you analysis anyway, it is an interesting start on questions that must be answered in the longer term so thank you.
If you flip your question, and ask which countries are having less number of patients and lower mortality due to COVID-19, you will find Bangladesh, India, Pakistan, Thailand, Malaysia still having less number of patients and lower mortality despite poor doctor to population ration and poor healthcare system compared to Germany, Spain, France, Italy, UK and USA.
SAR-CoV-2 (& COVID-19) has not read Medicine!
"You can not tell a disease how to behave". That is why if you look at the presentation of a disease it varies widely from patient to patient. Similarly, the countries having one of the best healthcare systems are helpless and hapless against COVID-19 while those countries with poor health care systems are spared e.g Africa, Indian subcontinent.
Muhammad Yousuf
"you will find Bangladesh, India, Pakistan, Thailand, Malaysia still having less number of patients and lower mortality despite poor doctor to population ration and poor healthcare system compared to Germany, Spain, France, Italy, UK and USA."
Yes, but one should not compare countries in very different stages of the evolution of the disease. India and Africa simply started later, it is much too early to state that they have lower mortality, even though one might hope so. In fact, India and Bangladesh already have lethality rates of 3.2%, which is about the same as in Germany, but when Germany had only 3300 cumulative infections as Bangladesh has now, their rate was near 1%. Pakistan is at 2.1%, which is pretty good with 9216 cumulative infected persons. But then this may be due to a higher dark figure due to lower testing rates than in India (which has about twice as many infections so far).
The Indonesian countries are different in that they have already had a long exposure to the disease but they managed to largely suppress it, being better prepared than Europe. They took action after SARS and the avian flu, which was essentially absent in Europe.
Hongkong has essentially managed to keep the disease small without a complete lockdown (they have 1030 cases and 4 deaths, which is super-low). They immediately reacted when it arrived (in January, when there were essentially no cases in Europe) and they enforced quarantines from the start and were able to track all contact networks. (Also they are disciplined in wearing masks and keeping up hygiene.) The illness never took off strongly in their city. Europe missed the chance to keep things at that low level, where you can still follow each single case. I am afraid that Africa will not fare better.
Here is an interesting hypothesis for part of the huge difference in lethality between countries such as Italy and Germany. It seems that the seriousness of the symptoms of a COVID-19 infection increases with the prior amount of exposure to air pollution. The regions with highest lethality in Italy are those of the industrialized north and they also have the highest air pollution. While this has decreased due to the breakdown of the industry (satellites are "seeing" that the air is now much cleaner in Northern Italy than it used to be in pre-Corona times), what appears important is the increased vulnerability of the immune system following long exposure to pollutants -- so it does not help much that the pollutants are reduced now, as patients have a wrecked immune system already.
Here are links to some of the relevant research:
Article Can atmospheric pollution be considered a co-factor in extre...
https://www.medrxiv.org/content/10.1101/2020.04.05.20054502v1The second article refers to research done in the US, where the same question is currently been investigated.
It would be interesting to have a similar comparison for different regions of Germany. We do have the highest numbers of infections in the three most industrialized regions of Germany, but I don't know anything about a comparison of the mortality rates. Of course, it would also have to be verified whether the environmentalist activists have managed to keep the air so much cleaner in Germany in general than in our neighbour countries to explain some of the differences. [I doubt it given how often we have "Feinstaubalarm" (fine dust alert) in some cities.]
Article Making informed decisions and balancing risk-benefit ratios ...
Please consider reviewing this brief letter to the BMJ.
It may provide some insight into healthcare in Bangladesh, a very poor developing country.
The current estimates of numbers of COVID-19 cases and deaths are very inaccurate. The sensitivity and specificity of the diagnostic tests for COVID-19 are not yet known in vivo. There are reports of very high sensitivity and specificity in vitro. However a number of confounding factors (e.g. the stage of illness and the way the swab was taken) mean that the false negative rate of these tests may be high in clinical practice.
Wenling Wang; Yanli Xu; Ruqin Gao; et al Detection of SARS-CoV-2 in Different Types of Clinical Specimens. Jama 2020 (in press).
The numbers of cases reported will also depend on strategies used for screening. Many countries are only screening symptomatic patients presenting to healthcare facilities. Others are attempting to screen a significant proportion of the population.
I believe the total number of cases of COVID-19 is significantly higher than that being reported. There are many confounding factors (both political and diagnostic) that make the initial number of deaths reported as being due to COVID-19 extremely difficult to interpret.
The availability of healthcare in general and specifically critical care will affect case fatality. In China and Italy where large numbers of healthcare professionals were infected the capacity to deliver healthcare was substantially reduced.
Another important factor to consider when trying to interpret variation in mortality rates between countries is differences in cultural interpretation of quality of life. This impacts the social and medical ethics and the delivery, limitation and withdrawal of life sustaining therapies.
In many Western cultures quality of life is valued far more than quantity of life. In those settings limitation and even withdrawal of life sustaining therapies resulting in cardiac arrest are a common mode of death.
In many Eastern cultures quantity of life is valued more than quality of life. In these settings treatment limitations and withdrawals of therapy are rare. Admission to intensive therapy units and prolonged organ support are common. Mode of death is often failure of resuscitation after cardiac arrest.
These differences are magnified when there are limited resources particularly in Western countries. The social and medical acceptance of treatment limitations is increased; this may significantly increase case fatality rates.
An unfortunate adverse effect of social distancing for the Covid-19 pandemic has arisen. There has been a dramatic fall in the number of blood donors in Saudi Arabia a result of social isolation during the Covid-19 Pandemic. Supplies of blood are being depleted. This is likely to be a global phenomenon
If insufficient blood of the correct type is available a woman may die as a result of not being transfused blood for severe acute bleeding following delivery of a child, for example. This will be most challenging for patients with rare blood group
Patients may die as a result of not being transfused blood severe acute bleeding due to stress ulceration, for example. If the patient has severe or critical Covid-19, the cause of death is actually the Covid-19 but it is likely to be coded as GI tract bleeding.
Rajkumar Rajendram
" The current estimates of numbers of COVID-19 cases and deaths are very inaccurate."
I agree regarding the number of cases. There is a high dark figure, because test frequency is insufficient in general and policies on testing are different in different countries.
As to the numbers of deaths, these should much more accurate. Of course, there can be misattributions. This will likely underestimate the number of deaths, as a death may be attributed to an accompanying illness rather than to COVID-19 itself.
Rajkumar Rajendram
"In many Western cultures quality of life is valued far more than quantity of life. In those settings limitation and even withdrawal of life sustaining therapies resulting in cardiac arrest are a common mode of death.
In many Eastern cultures quantity of life is valued more than quality of life. In these settings treatment limitations and withdrawals of therapy are rare. Admission to intensive therapy units and prolonged organ support are common. Mode of death is often failure of resuscitation after cardiac arrest.
These differences are magnified when there are limited resources particularly in Western countries. The social and medical acceptance of treatment limitations is increased; this may significantly increase case fatality rates."
There is an additional twist to this.
The Western approach also has an economic aspect. Limitation of life sustaining treatment (not really therapies, because they do not render you healthy again, otherwise no limitation would be imposed) liberates medical technology to be used on those that can be helped to regain their health.
So the limiting capacity will not be as easily reached as in Eastern cultures, where you need the care units to keep everybody alive, no matter whether they can continue living consciously or whether they are kept alive without consciousness.
So case fatality rates might also significantly increase in Eastern cultures due to overcharged capacities of intensive care.
In the end, fatality rates will tend to increase no matter whether the culture is Western or Eastern, albeit for different reasons.
I have compiled another table to get some idea about the "true" lethality rates. The most straightforward way to compute such a rate is to take the number of deaths occured so far and divide it by the cumulative number of infections. There are two major errors in this approach.
First, if one dies due to the disease, this will not happen immediately, but after an average duration of the illness. People who were identified as infected two days ago, will not normally die immediately but contribute to the denominator of the lethality rate calculated as described. So in reality it should be higher as only the group of persons infected sufficiently long ago can contribute. An enormous increase in infections would decrease the so-calculated lethality, which is not precisely what I would call intuitively true. Therefore, I have calculated a nominal lethality rate by dividing the number of deceased as of now by the number of infected persons until two weeks ago.
Second, there is a dark figure. Actually, more people are infected than are known to be infected, bcause only those detected (by tests or otherwise) are counted in. Taking the correct dark figure into account, one has to divide the nominal lethality rate by the dark figure to obtain the correct lethality rate (or case fatality rate). Now it is difficult to know the dark figure, so I tried to estimate it. There was a measurement of the case fatality rate in a scientific study of a particular community in Germany (Gangelt in the district of Heinsberg) via surveying a representative fraction of the population. In that region, one of the first outbreaks in Germany happened, and by now, the situation is essentially under control in the city of Heinsberg and surroundings. The measurement gave a case fatality rate of 0.37%, to be compared with the nominal rate for all of Germany of 1.98% calculated by the Johns-Hopkins university (which was probably obtained by dividing the current no. of deaths (around April 9) by the current cumulative no. of infected). Comparison of the two results would suggest a dark figure of around 5.35. Taking into account the current nominal rate of 4.8% as of 24 April and calculated with the number of infected persons 14 days before, we obtain a dark figure of 12, if we assume the Gangelt result to generalize to all of Germany and round it up to 0.4%. This requires an approximately homogeneous distribution of the case count throughout of Germany which is of course not given -- the Eastern part is (with the exception of Berlin) not as strongly affected as the Southern and Western parts. But this crude estimate will have to do as we have nothing better... The true dark figure may be a factor of 2 lower or a bit higher, but probably not by much.
To obtain the dark figure for other countries, I made the following assumptions.
1) It should be inversely proportional to the number of tests for the disease made in the country (these numbers can be found under https://de.wikipedia.org/wiki/COVID-19-Pandemie/Statistik#Anzahl_durchgef%C3%BChrte_Tests_auf_SARS-CoV-2_in_L%C3%A4ndern, the corresponding English Wikpedia page does not seem to give the number of tests by country).
2) Since it should be actually inversely proportional to the number of tests per fixed number of inhabitants, it should be directly proportional to the number of inhabitants of the country.
The second assumption is definitely not correct, if the disease is not yet spread (approximately homogenously) throughout the country. The first is doubtful during the initial outbreak of the disease, as then local tests may be sufficient to capture almost all cases, so fewer tests will not increase the dark figure.
If I make these assumptions, I find that the proportionality factor to get from the ratio population/number of tests = 83 million/2.07 million ~ 40 in Germany down to the 12 found from the nominal and measured percentages is 1/3.333=0.3. So I calculate the "true" lethality from the nominal one via multiplication with the population divided by the number of tests made and by 3.333. The results are given in the last column of the attached table.
First, it should be noted that the dark figures obtained are pretty high. For the US, we would have a factor of 20 between the found cases and the real number of infected persons, i.e. instead of 919 thousand, the number of infected persons would be 18.1 millions. Now, the so-estimated dark figure is not very accurate, it may well be a factor of 2 or 3 lower. But note that this would increase the corrected case fatality count by the same factor, so instead of 0.54% lethality we would have 1.08% or 1.62%. Also, for Iceland the dark figure cannot be a factor of 3 smaller, because they have such a small population and such a large number of tests that even now it is not above 2.3.
Second, some countries have enormous dark figures, either because they have made only a small number of tests (Iran) or have a huge population (India). It is clear that these are wrong and lead to incorrect lethality rates (which are far too low). For China, we would probably have a similarly low final rate (but they don't publish test numbers), because they also exceed a billion in population and it is clear that in this case, there is a huge inhomogeneity, with the disease being restricted to one region (Wuhan) with a population of about 10 million only, and tests mostly applied in that region, too. I would consider all estimates leading to dark figures above 50 as unrealistic in the table. So India will most likely have a true case fatality rate well above 0.02%, Malaysia one above 0.03%, and Thailand one above 0.00%.
The information that can be gathered from the table is that case fatality rates are between 0.2% and 0.7% for most countries and above 0.7% in those countries, where there was a partial breakdown of the health system (Italy, Spain, UK). The US escapes the latter mechanism, because New York, where the system did break down, is only part of the country.
These figures are very approximate and may be a factor of 2 or 3 larger in reality. Which shows that the lethality of COVID 19 is a bit higher than that of an ordinary flu, which would be around 0.1%.
Almost two months have gone by since I made my last list comparing the number of infections with COVID 19 and the case fatality rates in different countries. At the time, I sorted the list according to the number of cumulative infections. I did not cover all the countries listed by the Johns Hopkins university, because this exceeds my patience. This time, I did not sort the entries anew, even though Brazil, Russia and India have moved up into places 2 to 4, but kept the order of April 20, for better comparison between the tables. I have added Greece at the end of the list, because I find their interesting, how they deal with the situation. (Pretty good so far.)
Again, I have calculated a nominal case fatality rate by dividing the number of deceased persons as of now by the number of infected persons until two weeks ago, to correct for the fact that the outcome for newly infected ones is uncertain and therefore they should not appear in a ratio where they cannot possibly contribute both to the numerator and the denominator (they will not be dead for time reasons rather than for coping well with the disease).
Since I believe that my estimates of dark figures from the last calculation were grossly exaggerated in many cases (at least that is true if one believes the experts who think that the dark figure in Germany is below 2, and so it will be pretty low in other European countries as well), I took my old estimates and divided them by 4, except when they exceeded 50, in which case I just set them to 10. In the one case, where the dark figure was smaller than 4 (Iceland), I set it equal to 1. It might also be equal to 1 for New Zealand, which has declared itself Corona virus free (and indeed, their last confirmed new infection was on May 21, which is more than three weeks back. But they may have had unnoticed cases, otherwise their case fatality rate of 1.5% (in the absence of a dark figure) would be pretty high for a country that dealt so well with the disease as they did. (A few Asian countries coped even better: Vietnam could essentially stop the virus at 333 infections and has 0 deaths.)
Some countries have soared in their infection numbers since April 20, i.e. less than 2 months ago. Whereas in Spain the number of infections increased by less than 10%, the corresponding figure was about 20% for Germany. The UK and the USA roughly doubled their numbers. Russia multiplied the number of its infections by a factor of 7, India by a factor of 11 and Brazil by a factor of 14. The African country in the list, Ghana haes 8.5 times more cumulative infections than it had back in April. Seems the pandemic has reached Africa seriously.
Obviously, the number of deaths reported is not as solid a figure as one might think (no dark figure suspected...), because a certain creativity is possible in what you attribute to COVID 19 and what you don't. In Germany, the figure contains deaths by COVID 19 and "related" deaths. The US will report in a similarly honest way. That Russia has so few deaths was only at the beginning explicable by their enormous exponential growth of infections (growing so fast that deaths always lagged behind...). I suspect, they report only deaths diagnosed as due to the Corona virus, but not related deaths. Similar for Turkey. I simply doubt that their health system is so good that their uncorrected case fatality rate is 2% better than that of Germany. Within Germany, I noted that the figures for deaths due to the flu are wildly different according to different sources. There is the reported number of deaths due to the flu by diagnosis. This is well below 100 in good years (for all of Germany) and somewhat above 1000 in bad years. But the Robert-Koch institute also gives an estimate of excess mortality, i.e. deaths caused by the flu, but diagnosed as something else such as heart failure, etc. These are flu-related deaths and their number is below 1000 in good years and an estimated 20000 to 25000 in bad ones. So depending on how a country reports its COVID 19 deaths, either restricting itself to those diagnosed or else including Corona related deaths, the numbers may easily differ by a factor of 10. If you multiply the Russian death toll by a factor of 10, it is no longer impressive.
The virus has affected the rich and the poor without distinction. As for the treatment, it is certainly better to be in a state with a well-organized health system.
Matteo Nioi "The virus has affected the rich and the poor without distinction."
Seems that is not quite correct. https://www.theguardian.com/world/2020/may/20/black-americans-death-rate-covid-19-coronavirus
Black people in the US are poorer on the average that white ones. And they die more often from Covid-19. It does not seem that they are more susceptible, in principle, because in Africa, death rates seem to be generally lower than in Europe or the US. Of course, the counts reported from Africa may not be accurate, but I would at least not assume that there is a major difference in susceptibility to the illness based on race. So the difference in death rates in the US is more likely to be due to the different circumstances of life, among which we have different levels of wealth (also coming with different levels of health insurance, of course).
Probably I have expressed myself badly, however, the USA and Brazil, Europe and India have very different GDP and yet the prevalence is high regardless of per capita income ....
Matteo Nioi I agree with that. I don't know much about India. I think they first had a very strong lockdown and then could not keep this up due to protestations, which then resulted in many cases. In the US and in Brazil, clearly the governments did less than could have been done (in particular both presidents denied the danger for a long time), so they have major problems now. In Europe, some governments were slow which resulted in the health system being overburdened, others managed to avoid that -- the difference was about a factor of four in deaths. But many European countries are doing o.k. in the meantime. One of the countries that reacted most efficiently was New Zealand. They got the number of new infections down to zero in about a month's time, and the only cases they have now (on the order of 20) are visitors from outside...
K. Kassner It is misleading to try giving too simple answers to complex questions. The death toll is certainly not just a matter of "health system" but many variables come into play. I agree on the usefulness of the lockdown but the differences can also be partly explained by political choices, preparedness of the global system to deal with an epidemic, climate, genetics, population density and then certainly also by the type of therapeutic intervention. To understand which of the variables is more attributable to the difference given to the hand, a multivariate analysis should be done based on these and other characteristics. But surely the factors to consider are different ..
- Badges
- Science topic
- Similar topics
- Public Health
- Health