Model this asymmetric fatality rate?

The fatality rate in Wuhan is 4.1 percent and 2.8 percent in Hubei, compared to just 0.17 percent elsewhere in mainland China.

Is it really medical supply scarcity, as brought by the quarantine, as the NYT article seems to suggest?  Why is the fatality rate so asymmetrically distributed, or is it?


I’ve been wondering this too. One possibility is that we have relatively accurate counts of cases outside of Wuhan, but the true number of cases (including mild cases not tested) in Wuhan is much higher. If that were so, then the Wuhan fatality rate would look higher than elsewhere.

Some of the models I have seen online (the Lancet?) estimated something like 75000 cases in Wuhan by sometime in January. If true, that would explain a lot of the observed variation.

Another reason might be that Wuhan is where the outbreak started. I'd like to see weekly numbers. What is the mortality rate now compared to the start? Has the mortality rate fallen in Wuhan since China acknowledged that this was a problem? And as Alnair mentions below, maybe people in other provinces just haven't had time to die yet. With so few deaths (563), any reasonably range of mismeasurement in the prevalence of the flu is bound to have a large impact on measured mortality rates.

Dishonest reporting.

Heyy veгу cool blog!! Guy .. Excellent
.. Amazing .. І'll bookmark you blog аnd tqke the feeds аlso?
Ι am happy to seek out so many usеful infoгmation right herе in the post,
we wаnt develop more techniques ᧐n thiѕ regard, tһanks foor sharing.
. . . . .

I suppose the obvious answer is that the virus is adapting to the early deaths of some host organisms. Milder forms of the virus may gain an evolutionary advantage by keeping hosts live longer to increase the opportunity for propagation.

I came here to say what JFA said, but JFA already said it.

The other aspect, perhaps related to supplies, is identification of infections. The deaths are much easier to see and then to identify (Zero Hedge aside) that mild and less sever cases. As such, one might expect a bias in the ratios to higher incident in Wuhan than outside the province of country.

You're right. They don't have enough test kits. If you have too few test kits, it makes no sense to test everyone in Wuhan/Hubei. These people are in quarantine anyway, so there is no urgent need to test all people with mild symptoms.

But of course you want to test everyone with mild symptoms outside of Wuhan/Hubei as soon as possible, because they are rather few and because those cases are really important to prevent the virus from spreading.

Massive under-reporting of cases? If it's the case, that would be actually not so bad. It would mean that there are so many cases that containment is useless and, since the fatality rate is low, we can focus on cure.

Another possibility is that people who are infected in Wuhan and Hubei are systematically different. If the disease started in Wuhan, it would affect potential anyone but people who are affected outside Wuhan are mainly people who traveled there. If travelers are younger and in better health, they may be less likely to die.

"It would mean that there are so many cases that containment is useless and, since the fatality rate is low, we can focus on cure."

Containment slows the spread. Slowing the spread is vital even if you fail at 100% containment.

Perhaps a 3rd. If the infections did originate from wild meat perhaps the initial infection is more detrimental to humans. As the infection is passed through multiple human immune systems -- 2nd, 3rd, 4th hand infections -- the mutations actually result in a less sever strain of the virus.

Because the numbers are meaningless. The number of dead are easy to count, and have to correspond roughly to people's experience. But the number of sick? That is a political construct. Dividing the first into the second and expecting it to have any connection to reality is ridiculous.

Agreed. See, e.g.,

Though unfortunately there are also (anecdotal, but necessarily so) of people dying at home in Wuhan, and thus not being "counted".

Could be:
- Coronavirus need time to kill, so new regions do will eventually reach same fatality rate.
- Many Infected had survived but not identified in first stages of the outbreak

This is what I’d bet on. What you really want to measure is the ultimate fatality rate of a cohort of people exposed at approximately the same time.
At the very least, I think it is safe to say the difference is caused by issues in counting and estimating and not in genetics or medical competence.

The death rate ought to fall with time. The first cases were likely confused with flu and recent cases are treated much more aggressively.

Even outside the medical system, if you suddenly came down with a severe cough today, what would you do? Is it any different from what you would have done three months ago? Do you think you would be a little quicker to isolate yourself from coworkers and maybe call the doctor?

There is a leak from Tencent outlining more believable figures (albeit "unofficial"

Nice, I almost thought that article was from CNN, a mainstream news source. The first problem with that article (besides it being hearsay from a Taiwanese paper that clearly has a political axe to grind if you look at their other articles) is why would Tencent a random tech company have this data? Perhaps it is actually data from Wechat rumors or something? Moreover, if the mortality rate was anything close to 16%, far more of the hundreds of foreign cases would have died, but in fact only one has.

It’s entirely believable that the Chinese government reacted in a draconian way to a statistically small but highly publicized problem. That’s what makes them an authoritarian government. It’s the same reason they detained a million Uighurs over a few hundred deaths in Uighur terrorism (perhaps they should also try to angle, the real number of Uighur terrorism deaths is 100 times the reported number too?)

Zaua asks " why would Tencent a random tech company have this data? "
Tencent is the company tasked by China to report out the official numbers. Therefore, they have the data. They reported out these numbers and then said it was a mistake. They did it twice. Since China has been denying the there was a virus for over a month, people in South East Asia are understandably skeptical and believe that it is an attempt to leak out the truth. Because the fact that we even know that a virus exists is a result of leaks through social media. The fact that the Western press treats Chinese naively is the problem.
Zaua writes, "It’s entirely believable that the Chinese government reacted in a draconian way to a statistically small but highly publicized problem. " No, this is not believable. The Chinese government was denying that a virus existed up until a week before it suddenly quarantined 65 million people, shut down one of its biggest cities and announced that it would build a hospital in a matter of weeks. It is not believable that the Communists would cause themselves massive economic harm because a mild virus is spreading. They don't care about the health and welfare of their citizens generally. Why now?

If the government had real numbers that were dozens of times the reported numbers, why would they entrust the real numbers to Tencent and risk this kind of error? Moreover, if it happened once, you’d think an authoritarian government trying to do a coverup on that scale would stop giving data to Tencent and maybe punish the leaders of Tencent for propagating this conspiracy theory. This theory is Infowars levels of plausibility.

The quick transition from the cover-up to the mass quarantine is easily explained by the fact that the local Wuhan government was in charge at the beginning, and they wanted to cover it up to prevent their city from being quarantined. Later, the national government took over and because they (as you say) don’t care about the health and welfare of Wuhan citizens, they were willing to enact this extremely draconian and apparently counterproductive quarantine.

"The quick transition from the cover-up to the mass quarantine is easily explained by the fact that the local Wuhan government was in charge at the beginning, and they wanted to cover it up to prevent their city from being quarantined." You don't know what you are talking about. The Communist party is in charge. There is no "local government" as if China is federated. The numbers coming out of China are garbage. A Hong Kong professor has warned that the official numbers only include those admitted to a hospital, but in Wuhan the hospitals were overwhelmed and were turning people away. There are social media clips of people collapsing in the streets of Wuhan. Social media in Hong Kong has been talking about the virus since December, but when China admitted that the virus existed, they said the first case was December 31. Now, scientists have confirmed that the virus was spreading human to human in early to mid December. There are reports that China is burning bodies to hide the extent of the death toll. China is now threatening to execute anyone who spreads "fake news" about the virus. It is hard to know from outside what is actually happening, but we have one source, China, that we can confirm has already lied to us, and on the other hand, we have leaks out of China through social media, that have been proven true. Here is another, it ought to be taken seriously.

The central government isn’t in charge of running the healthcare and public health systems at the provincial level. Of course the central government can order around anyone however they want, but in practice the provinces are actually running a lot of things because the central party doesn’t have the manpower to micromanage every aspect of the country. Just like any other large, bureaucratic organization.

Of course, Xi doesn't putting in the order for more antibiotics or surgical masks. What is implausible is that the local government on its own hid the existence of the virus, and Xi's government issued denials out of ignorance and then reversed course once it learned the truth, and then completely over-reacted. They could have, but what is more likely is that they hid this because they hid all negative news, and then they lost control of the situation, and they are still lying.

I find it very plausible that local authorities tried to hide this from the central party for at least a little while. But I also find it implausible that the local party is solely responsible for the delays in reporting this to the Chinese public or to the world.

Zaua writes: "Moreover, if it happened once, you’d think an authoritarian government trying to do a coverup on that scale would stop giving data to Tencent and maybe punish the leaders of Tencent for propagating this conspiracy theory. This theory is Infowars levels of plausibility." Conspiracy theory may be implausible in the West where there is no ability to control every source of information, but China does in fact control every source of information. Only Chinese with there own VPNs can get information out, and that under threat of death. So, no this is not an Infowars level conspiracy. The Chinese Communist party actually operates concentration camps, kills dissidents, forces anyone who does business with them to accept their propaganda. As far as Tencent, it is a Chinese company. It is not independent from the Communist party. It is controlled by them. The government was using it to disseminate data. I am sure that the Communist are searching for the person who leaked this data, and hopefully he hid his tracks. But, as you say, it is also possible, that the Communist party had been denying the existence of a virus for over a month because the central government was deferring to the local goverment who was lying to Beijing and Xi about the extent of the virus. It is possible that those local officials, who apparently had no fear of lying to Xi, finally came forward and told Xi that they had lost control of the situation, but that the virus still was that contagious. Then, Xi decided to risk throwing China into recession by quarantining 65 million people even though his scientists would have certainly told him that such actions are not necessary for such a relatively mild virus given the official numbers. Xi then also decided to build a hospital in two weeks even though the Wuhan medical system was not completely overwhelmed. All of that is possible. But, what is more likely is that China is lying. That the Central government has always been calling the shots and that the only reason we know about any of this is because people in Wuhan were so desperate that they were willing to defy there government because they are dying anyway.

"Since China has been denying the there was a virus for over a month, people in South East Asia are understandably skeptical and believe that it is an attempt to leak out the truth."

This is certainly true in Thailand where local language papers have promoted this story.

We'll know that those Tencent numbers are real when orbiting satellites see bulldozers digging plague pits.

Maybe we already have seen it.

A hospital with a thousand beds built in a couple weeks.

Charnel house disguised as hospital?

I have no idea if the Tencent numbers are correct, but they aren't high enough that you would require plague pits. They are reporting 25K dead. Even if that was nearly all in the province of Hubei, it's still a population of 59 million.

General mortality rate in China appears to be 7.13 per 1,000. That's about 1,200 per day on a normal day. If the 25K were all in the last 10 days, that would alone would only be twice the normal death rate.

I'm not saying this is good, nor am I claiming the numbers are accurate, I'm just pointing out that even if they are, they aren't catastrophic.

It could be food and medical treatment (not just a lack of supplies, I’m seeing reports that people can’t even get to the hospital because public transportation is shut down and taxis have been commandeered by the government—most people in China are not wealthy enough to have private cars), or it could be that the younger and healthier people were more able to flee leaving older and more likely to die people in Wuhan and Hubei.

If this is a data reporting issue though, that would probably be a good sign because under-reporting is more likely occurring in Hubei province where people are overwhelmed than elsewhere in China. Under-reporting deaths in Hubei would mean the asymmetry is even larger and harder to explain than in the data. The alternative of under-reporting cases (perhaps because people with milder cases can’t get to the hospital) seems the most plausible, but also means the true mortality rate is lower than the Hubei numbers and perhaps closer to the Mainland numbers. There aren’t enough foreign cases yet to draw reasonable statistical conclusions and it will hopefully stay that way.

But at least this sort of data will be very valuable in assessing how to respond to future outbreaks, and when quarantines can do more harm than good even judged solely on the criterion of disease deaths.

Would migrant workers be likelier to be younger and healthier than people with Wuhan residence cards? It would seem to me that the likeliest people to have fled would have been people who have family elsewhere.

Surely it's the opposite of scarcity in the rest of China, as brought by the quarantine? The whole point of the quarantine is to delay the spread so other places get to deal with the disease spread out over time, without clogged hospitals and sick doctors, with access to stockpiled antivirals etc. To say nothing of the positive effect of time on our ability to find effective treatments and a vaccine.

Delay the spread to spring and summer and the epidemic will be a lot less problematic.

Wouldn't this be explained by the cases in Wuhan being more progressed than elsewhere, with a lag in mortality?

+1, this is almost certainly the answer. And it's what the China CDC has actually reported.

Of course, people need to stop with the "it doesn't seem so bad, let's not panic" crap. We don't know. We cannot trust the data coming out of China, and there simply are not enough cases far enough progressed outside of China to know how deadly the virus is. We'll know in a couple of weeks. Hopefully, this isn't so bad but China is acting like this is a major catastrophe while telling the West that it isn't that bad.

"Let's not panic" is almost always good advice.

What "Don't panic" is is condescending response to people who are raising concerns that they are not being given complete and accurate information? It is a way for those in power to make those asking difficult questions look like nuts, so that they can continue to lie.

You are not asking difficult questions. You are looking for reasons to be afraid. There is huge demand for this in the 21st century. Just stay out of Hubei and you will almost certainly be fine.

Care to schedule any trips to Africa, India, or Cambodia this spring?

Are Chinese authorities at this point in time saying that the situation isn’t that bad? My impression is that the Chinese government is mostly trying to do what it can to preserve its image for the Chinese public, which means a mixture of trying to look they are acting effectively to control the spread of N-Cov, standard nationalist propaganda, and some reassurance that they have things under control, and that the folks saying “Don’t panic” are mostly Westerners. I think that the Chinese government wants the Chinese public to be worried about the virus and reassured by the government’s response.

5 million left Wuhan after the outbreak and before the lockdown, those leaving likely young and healthy, those left behind likely older and less healthy. China is attempting to round up those who left, demanding that they be identified by families, friends, and neighbors.

From today's WP online:

"China began clinical trials in the virus-hit city of Wuhan to determine whether coronavirus patients can be treated with the antiviral drug remdesivir. Professor Wang Chen and Cao Bin, pulmonologists from the China-Japan Friendship Hospital in Beijing, said Wednesday that they would study 761 adult patients in two trials — one to assess 453 severely infected patients, the other on 308 with mild or moderate symptoms — at Wuhan Jinyintan Hospital.

“We anticipate good results to be achieved in the clinical trials,” Sun Yanrong, deputy director of the China National Center for Biotechnology Development under the Ministry of Science and Technology, told a news conference in Beijing earlier this week.

"Gilead Sciences, the U.S. company that originally developed remdesivir for Ebola and Marburg virus infections, is offering the drug free to support the Chinese trials, which are expected to be finished on April 27. Remdesivir has not been approved for use by the Food and Drug Administration, although doctors can pursue “compassionate use” of a drug in the event of a serious health condition.

"The Wuhan Institute of Virology has applied to patent the use of Gilead’s remdesivir to treat the current coronavirus outbreak, a move that could spark fresh tension with the United States over intellectual property and technology transfer. Gilead’s patent application, filed in China in 2016 before the virus was identified, cites only the overall family of coronaviruses. “Gilead has no influence over whether a patent office issues a patent to the Chinese researchers,” company spokesman, Ryan McKeel told the Associated Press. “Their application has been filed more than three years after Gilead’s filing and will be considered in view of what is already known about the compound and pending patent applications.”

"The Chinese researchers said they had made their patent application “from the perspective of protecting national interests,” the institute said in a statement. “If relevant foreign companies plan to contribute to China’s epidemic prevention and control, we both agree that if the state needs it, we will not require enforcement of rights given by the patent,” it said.

"Is it really medical supply scarcity, as brought by the quarantine, as the NYT article seems to suggest? Why is the fatality rate so asymmetrically distributed, or is it?"

Tyler, from the other thread:
Per the China CDC:

0.16% of infections outside of Hubei have died so far but they expect the number to grow to 2.1% as patients that are in early stages die.

Mortality rate in Wuhan was 4.9%.
Mortality rate in the Hubei Province was 3.1%.
Mortality rate nationwide was 2.1%.
Fatality rate in other provinces is 0.16%.

This link talks about the issue. The best estimate by the Chinese is that the Case Fatality Rate will be 2.1%. It was much higher in Hubei do to surprise and a lack of medical supplies and facilities.

You need to wait for a huge number of samples until the law of large numbers kicks in. Wuhan is ground zero so it gets their faster than anybody else.

+1 Also, people are acting like the death rate is an inherent property of a virus. The virus' death rate will vary depending on all sorts of factors. We should expect the death rate to rise or fall depending in which population it spreads, the response of the healthcare system. All this worthless speculation just risks giving people a false sense of confidence.

The data is probably fictional so modelling it is foolish (unless you are a Climate Scientist).

I do beg your pardon. The data are probably fictional so modelling them is foolish ...

Is it foolish? it certainly isn't a waste of time, modeling it. I think everyone understands the data and informations is suspect; it'd be foolish to take it at face value. But similar to climate change, it's probably far better to take it very seriously with the potential of over-reacting than to stick one's head in the sand until the data is irrefutable. Or, I guess in both cases, as some seem to prefer it, humans can wait for barn animals to solve both problems; since it's their problem too.

Begging various pardons, as well...

Don't trust those numbers. Statistics is the one area of math that the Chinese aren't very good at.

Viruses mutate pretty quickly, and greater contagion tends to go with lower morality. It is possible that a less fatal variant has emerged and is outpacing the original strain already.

That's MORTALITY. Autocorrect, you so funny...

An alternative reason maybe the disease is infecting everyone in Wuhan and the people outside Wuhan were young and healthy enough to travel to Wuhan to be infected.

You yourself linked to a blog explaining that not a single data point from China is to be trusted. Right?

There is good reason to believe the Iowa results have been manipulated by the Buttilieg wing of the Democratic Party:
We can't let Wall Street steal another election.

Hard to make meaningful conclusions when your data is likely nonesense

What constitutes an infection is clearly on a sliding scale based on where people are. We know that for many people, perhaps most, an infection is asymptomatic. For many others the symptoms are mild. The measurement used for the denominator is highly suspect. Using the total population of a region as the denominator would produce a more useful ratio, albeit one that won't tell the full story until everyone has had their chance to get infected and die.

Wuhan is a metropolis, it is the capital of the Hubei province. The more urban something is in China, the worse the air quality. [citation needed] Wuhan is five times larger than the next biggest city XiangYang (7 million vs 1.5 million). Smog isn't conductive for living long and prospering (or surviving a mean virus). What was the mystery here again?

Ah right.... the divide between Hubei and the rest of China. Well, since it started in Wuhan, the infected that would die to Corona just ain't dead yet. And Wuhan being first had the least time to prepare for the Corona virus. So yeah.... scarcity of supplies makes sense. Also I'd look at an air pollution map overlaid with number of infections for "the rest of China". Average air quality for someone infected outside of Hubei would be interesting.

Guangdong Province in southern China has the largest number of cases outside of Hubei, and also has the cleanest air in China due a relative paucity of coal power plants and heavy industry. The province gets its electricity from the Three Gorges Dam and also from a number of nuclear reactors. It is also more focused on light industry and electronics assembly (and IT in general). Now, maybe there will be a lower case fatality rate there. Some people have already speculated that smoking makes it likelier that someone who catches the virus will get a severe case, as Chinese media keeps reporting that older folks, men, and people with underlying health problems are at most risk. But if you have been to China you know that most men smoke and very few women do.

The cleaner the air, the more people will spend time outside. That might be a factor for more widespread transmission.
Trouble is, that if the air is clean, more vulnerable people would also spend time outside and get infected, so it might just end up being a wash.
If the fatality rate turns out to be significantly less for areas with less pollution though, maybe it'll have an impact on environmental policy.

"Well, since it started in Wuhan, the infected that would die to Corona just ain't dead yet. And Wuhan being first had the least time to prepare for the Corona virus"

+1, this is essentially what the China CDC has reported. I don't really understand the confusion. Other than obviously the reporters at the NYT's weren't exactly the smartest kids in their class.

It seems that for a good number of cases (~25%) extensive care is required due to relatively sudden onset of severe symptoms after about a week of milder symptoms. That seems to be a pattern with the first case, well documented, in the US and also the cases in Italy and elsewhere.

If that's true, regions with a small number of cases will be able to have resources to provide intensive care to small number of cases, but will become overwhelmed once the case number exceeds several hundred. Then more people will die, particularly those with pre-existing conditions who are unable to obtain necessary ICU care.

Thus, the fatality is likely regionally determined by the capacity of the regional health care resources.

We also know that testing is not trivial or fast, but have evidence of rapid spread in isolated clusters where testing is universal. The Japan cruise ship shows one instance of rapid spread of the virus to 20 passengers in short time. The German case at Webasto also shows rapid spread to ten cases within a short time. All these cases are not at the severe stage and thus hard to miss without mandatory testing for all people in a possible infection cluster.

I'm worried about less developed countries with extensive trade-links to China along the Belt and Road Corridor. I find it personally impossible that Japan, Singapore, Australia, and South Korea are among the hardest hit countries, but that there are no cases in Pakistan, Bangladesh, all of Africa, and just three cases in India. That must be due to lack of detection.

"Thus, the fatality is likely regionally determined by the capacity of the regional health care resources."

The availability of respirators seems critical. That may be the single biggest reason that Wuhan had a Mortality rate twice as high as the China CDC believes is warranted. The existing hospitals quickly filled up and there weren't enough respirators to handle the patients that required them.

So far the majority of deaths are in groups you would expect. elderly males with co-morbidity risks. 2/3 are male. 80% over 60. 75% have an underlying disease. The number of critical care beds is very limited and they seem to do a very bad job with complex cases with co-morbidity risks. Simply they don't seem to do well with seriously ill elderly patients, nor do they have the infrastructure to deal with these patients. Wuhan has a population of over 8 million that has grown by about 25% in the last twenty years.

Many national health systems are good at routine care. They tend to do less well with complex cases.

The average age of the infected outside China is 45 with 71% male. This may help explain the lower fatality rate.

On the downside, it appears to be very contagious. The average infected person spreads it to 4.08 other people. the flu is typically 1.3. SARS was 2.0. So far only about 20% of infected people are considered " severe".

One morning headline begins to raise another question concerning mortality rates:

how many people begin to die after poultry and other livestock themselves begin succumbing to "the Wuhan coronavirus"?

Once diseased poultry and livestock become inedible, how much tofu is left to go around?

Good time to view Bong Joon Ho's The Host.

WHO whoa: is the Situation Report for 6 Feb 2020 tardy in being posted? Looks like it's a couple of hours late already . . . don't tell us WHO is using an ACRONYM/Shadow app, please.

Once for a while, I worked for an institute where Dr. Ho Mei-Shang's lab resides, so I often saw her new posts on the walls. She is the SARS expert in Taiwan.

One of many discoveries she made in 2003-2004 is, there is a 23-day cycle with which the surface cells of lungs are refreshed by stem cells, and SARS virus is good at killing these lung stem cells. Once all such stem cells of a patient are killed, even if the lung surface looks good and new, the patient will be in a fatal state once the 23-day cycle is due, because there is no new cell to replace the deprecating self-killing old surface cells.

If the theory sticks with the novel coronavirus this time, there is high chance that many patients who have pneumonia symptom would be killed on the due time of their 23-day cycles. It is just not yet the time for them to die.

If the 23-day cycle theory applies, newly infected patients outside Wuhan would significantly lower the average fatality rate for one or two weeks, before those unlucky ones of them beginning to die.

The more intelligent workers can work from homes. Thus on average the higher the smart faction for the province the lower the provincial fatality rate. However only the IQ120+ smart fraction is statistically significant and negatively correlated with the fatality rate. Nevertheless, the data for ground zero Hubei are too high and if they are included it will overwhelm the results to be insignificant. The economy of Hubei are more on heavy industries and high tech components manufacturing where most of the employees cannot work from homes, not like the financial and software development sectors of Shanghai, Beijing, Anhui and Zhejiang.

Comments for this post are closed