|
China Did Not Deceive Us – Counting Death During An Epidemic Is Really Difficult
The anti-China campaign, which the Deputy National Security Advisor Matthew Pottinger is running, presented its April fools joke. It leaked to Bloomberg that a secret U.S. Intelligence Report claims that China concealed the real numbers of its Covid-19 cases:
China has concealed the extent of the coronavirus outbreak in its country, under-reporting both total cases and deaths it’s suffered from the disease, the U.S. intelligence community concluded in a classified report to the White House, according to three U.S. officials.
…
While China eventually imposed a strict lockdown beyond those of less autocratic nations, there has been considerable skepticism of China’s reported numbers, both outside and within the country. The Chinese government has repeatedly revised its methodology for counting cases, for weeks excluding people without symptoms entirely, and only on Tuesday added more than 1,500 asymptomatic cases to its total.
Stacks of thousands of urns outside funeral homes in Hubei province have driven public doubt in Beijing’s reporting.
China did not conceal its number of Covid-19 cases. Nor did it hold back any information.
Reporting numbers during an outbreak of a new disease is actually very difficult.
When does one start to count? China only knew that it had a new virus epidemic in early January. By then those who died during the month before were already cremated. How could it count them?
Does one include co-morbids or not in the count? What about casualties of a car accident that also test positive for Covid-19 when they die? What about those who died with Covid-19 symptoms but could not be tested for lack of test kits? Are the tests really working reliably? At one point China included all pneumonia cases in the Covid-19 case count even after they tested negative for Covid-19. The Chinese epidemiologists thought that their test had been wrong and only later found out that that was not the case.
What about asymptomatic cases that test positive. Are these false positives or do these people really have the virus? One can only know that by testing them a month later for antibodies. If they developed antibody cells against the virus they must have had it. That may well be the reason why China only now added the 1.500 asymptomatic cases to its total count.
The most important number during an outbreak is the one that lets one plan for resources and model for countermeasures. That number is the Case Fatality Rate.
bigger
But that is the wrong number if one asks how likely one is to die of the disease:
You may have heard a term being used: the “case fatality rate”, or CFR. That is the number of deaths divided by the number of confirmed cases. When journalists talk about the “death rate”, that’s often what they are referring to. If a country has 10,000 confirmed cases and 100 deaths, then the CFR in that country is (100/10,000), or 1%.
That is not what we are looking for, and it is probably not even very close to what we are looking for.
Instead what we want is the “infection fatality rate”, or IFR. That is the number of deaths divided by the number of people who actually have the disease. The number of people who have tested positive for the disease is probably only a fraction of the total number who had it, because only a fraction of the population has actually been tested.
Obviously, the IFR is much harder to determine accurately. The only people getting tested will be the people who are most ill, so your IFR is probably much lower than your CFR, because your denominator — the number you’re dividing by — is probably much bigger.
So if your country has tested absolutely everyone and found all cases of the disease, then your IFR is the same as your CFR, or 1%. But if it has only found 10% of the people with the disease, then your 10,000 confirmed cases are just the tip of a 100,000-person iceberg. With those 100 deaths, your IFR would be (100/100,000) or 0.1%.
China, and everyone who followed its data, knew that the number of cases were different from the number of infections. But we did not know by how much. It was also clear that China was not counting all Covid-19 death. Italy shows how that problem arises:
As hospitals become overcrowded, patients are being asked to stay at home until they display the most serious symptoms. Many will die in their houses or nursing homes and may not even be counted as Covid-19 cases unless they’re tested post-mortem.
Last week, two researchers from northern Italy made this point forcefully when looking at Nembro, a small town near Bergamo that has been very severely hit by the outbreak. Writing in Italian newspaper Corriere della Sera they found there had been 158 deaths in the town in 2020 so far, as opposed to 35 on average in the previous five years. They noted that Nembro had only counted 31 deaths from Covid-19, which looks like an underestimate.
In other towns nearby, including Bergamo itself, the trend seemed identical. The researchers made the point that the only reliable indicator in the end will be “excess deaths” — namely, how many more people have died in total compared to a “normal” year.
Death per month in Bergamo over the last ten years*
*The data refer to deaths until March 26th
Source – bigger
The UK produces two different numbers. The Office for National Statistics says that it counts more Covid-19 death than the official GOV.UK site by the Department of Health and Social Care:
- We include all deaths where COVID-19 was mentioned on the death certificate, even if only suspected: the GOV.UK figures are only those deaths where the patient had a positive test result
- We include deaths that happened anywhere in England and Wales, for example some might be in care homes: the GOV.UK figures are only those that happened in hospital.
The definition who to count can change over time and not only in China:
[C]ountries may have good reasons to change the way they collect data as circumstances change, but it apparently happens often enough that the World Health Organisation feels that they have to ask countries to notify them when they do it. Famously, China did so earlier in the epidemic, but others do too: in complying with the WHO’s request, Australia has noted that it has changed its definition of a Covid-19 “case” (and therefore a Covid-19 “death”) at least 12 times since 23 January.
As for the number of urns delivered to funeral homes in Hubei after the quarantine was lifted one has also to consider the number of regular death. Hubei province has some sixty million inhabitants. The regular mortality rate in China is 726 per 100.000 inhabitants per year. The regular expected number of death from January 1 to March 31 in Hubei province without the epidemic was 108.900. In Wuhan, which has 14 million inhabitants, the expected number was 25.410. Photos that show the delivery of a few thousands of urns to large funeral homes in Wuhan are thereby not a sign for a higher Covid-19 death rate. To claim such is propaganda nonsense.
There is no reason to criticize China for publishing incomplete and a times confusing numbers. That is normal during any epidemic and the U.S. will certainly do likewise. The real problem with the various numbers flowing around lies elsewhere.
People do try to make predictions about how many will get infected and die from the virus. These models are needed to prepare ones resources. But prediction is extremely difficult to do as the various models are very sensitive to the input data. A model that works in country A may give the wrong results when it is used for country B. Cities and towns are different. Local circumstances can make huge differences. With the real infection numbers and the real death rate unknown during an outbreak we can only hope that our epidemiologists, who are trained to make and interpret such models, get it right.
To claim that China deceived the U.S. and the world about its numbers or that China tried to make it look as if the epidemic was not as serious as it is makes no sense at all.
China took extreme and drastic measures at high economic costs to prevent a larger outbreak. It did not do that to deceive anyone but because it saw the seriousness of the problem. It acted in the global interest and to defeat the virus.
China gave the world time to prepare for the pandemic. Unfortunately that time was not used well. One reason that the U.S. will now experience a very large outbreak is that it is not willing to follow China’s example. If one declares that gun shops and shooting ranges are critical businesses that must stay open during a lockdown one is not serious about fighting the epidemic.
To blame China for that is simply nonsensical.
The real number of casualties the SARS-CoV-19 outbreak will cause will only be known when it is over and when we compare the new death statistics to those of previous years. One thing is assured. The “excess death” numbers will be lower in those countries that did use the time China gave them and prepared for what was coming at them.
—
Previous Moon of Alabama posts on the issue:
We’re all epidemiologists now.
Here are my yet unformed opinions and questions on CoV-2. Ignore this if you have coronavirus fatigue!
The Case Fatality Rate (CFR = Number of deaths/Number of confirmed cases) is 11% of in Italy, 7% in Spain, 4.7% globally, 1.7% in South Korea, 1% on the Diamond Princess, 0.06% in Germany, 0.06% in Iceland and 0% in the city of Vo in Italy. What’s going on?
Are the big differences due to variations in government intervention timing, or rather in data methodologies and reliability? Is the coronavirus growth really exponential, or more similar to typical flu season that has an early exponential stage, then peaks, and then declines without state intervention? It’s easy to underestimate the growth in the early linear growth phase, and overestimate it in the later exponential phase of any epidemic. The exponential phase is not sustained indefinitely even without interventions, though is affected by them. COVID-19 seems to have been sensitive to interventions (social distancing etc), but we don’t know for sure the size of this effect.
https://voxeu.org/article/it-s-not-exponential-economist-s-view-epidemiological-curve
Questions like these need good data, and in particular randomised testing of the general population. This informs other data like the overall death rate, how to best reorganise society, and forecasting. Infection Fatality Rate (IFR) is just as, or more important than, CFR. IFR = Number of Deaths/Number of infected. If the infection rate is actually twice what we previously thought, then the IFR is half. As the denominator in this ratio increases because more people are diagnosed with the infection, then the percentage of death decreases. We need to know the denominator reliably, and that’s the missing part of the puzzle. This affects what kind of response we make. Data is missing mostly due to the limited number of test kits, as most are currently triaged towards active cases, so the important data that we need to make consequential, society-wide decisions is missing.
But which test kits? The antigen PRC test using a nasal swab, shows whether a person is infected, through direct detection. They have a reputation some false positives and up to 30% false negatives. The antibody test using a finger-prick blood sample, shows whether a person was infected, via their immune response. These kits take longer to manufacture, and their accuracy is not yet known. We need both kinds of tests to gauge the asymptomatic, illness and death rates, and make important comparisons. They are in the pipeline.
https://www.ft.com/content/0faf8e7a-d966-44a5-b4ee-8213841da688
The following may give some indicative data points for the most controlled populations to date:
The Diamond Princess tested everybody and had a 20% infection rate, and a 1% fatality rate. It was an aged cohort, so the death rate may be lower in a national population.
Iceland tested 4.2% of its population, both sick and healthy people who volunteered (not quite randomised but better than testing only the symptomatic), and found a 0.9% community infection rate, and a 0.06% fatality rate.
The Italian city of Vo tested everybody, and found a 3% community infection rate. Because the infected were identifiable, separating the sick from the well was effective, resulting in a zero death rate.
Seasonal flu in the US has about 35 million symptomatic cases typically with 35,000 deaths p.a., but in a bad year can be much more e.g. 59,000 in 2018. The current number of US coronavirus deaths is 3,700. New York is now the epicentre in the US, but I can’t find data on community infection rates, or symptomatic case rates. The current corona death rate is nearly 1,000, while the 2018 seasonal flu death rate for NYC was 4,749.
Each year, seasonal flu makes between 3% and 11% of the entire US population sick. The number of people who get infected, including non-symptomatic cases, can only be estimated because randomised testing is not carried out. It’s thought to be between 5% and 20%.
https://www.cdc.gov/flu/about/burden/index.html
https://www.nbcnewyork.com/news/local/as-nyc-nears-1000-covid-19-deaths-how-does-it-compare-to-typical-flu-seasons/2352180/
https://academic.oup.com/cid/article/66/10/1511/4682599
Seasonal flu in Italy between 2014-17 resulted between 7,000 to 25,000 deaths p.a. … much higher than most countries. There were many contributing factors to its high coronavirus death rate e.g. early panic reducing their triage quality, high-age demographics (average fatality age 81 years), high percentage of smokers, relatively now number of ICU beds, and routinely 90% full, meaning little spare capacity. Note that their 10,000 corona deaths are still at the lower end of their seasonal flu death range.
https://www.sciencedirect.com/science/article/pii/S1201971219303285
The Italian city Bergamo is an outlier. But about 10,000 people (1/3 of Bergamo’s population) attended a Champion’s League football match against Spain in Milan on 19 February, then partied on public transport returning home, centimetres apart. Others watched the match close together in bars and homes in Bergamo. Together this created a “biological bomb”. What is the community infection rate in Bergamo? 5%? 10%? 50%? We don’t know.
https://www.independent.co.uk/news/world/coronavirus-italy-champions-league-atlanta-valencia-milan-bergamo-a9426616.html
South Korea is an interesting example. They tested only 0.8% of its population, but employed mostly test-and-track strategy, which was effective. However, I don’t think they did randomised testing of the broad population. If so, it’s 1.6% death rate is based only on confirmed symptomatic cases, not community-wide infections.
My instinct is that Covid-19 is more infectious than other coronaviruses (maybe because it’s airbourne, rather than just droplet transmission?), but with a lethality on a par with seasonal flu, except for the co-morbid aged for which it is sharply higher. There seems to be a higher death rate in the extreme cases, though we don’t know. Seasonal flu kills between 300,000 and 650,000 people per year (WHO). The Spanish Flu (which actually started in the US) was most lethal towards healthy young adults due to their stronger adverse immune reactions. Swine flu (which also started in the US) did not target the aged, and killed between 150,000 and 575,000 globally in its first year, according to the CDC. (How’s that for an error range?) But without good data we don’t know. Coronavirus scares me. But what is happening to our society right now really scares me. An economic depression, unemployment, poverty, spiking inequality, hunger, domestic violence and an overbearing police state could result in a higher death rate than coronavirus itself, if we are not wise. Hopefully the current lockdown is short, sharp, and effective, and buys us time to do those tests then make intelligent, informed, public-health decisions. Very soon.
https://www.pnas.org/content/115/5/1081
I’m interested if you have evidence that contradicts this analysis. The devil is in the data.
I found Professor John Ioannides of Stanford useful. https://www.youtube.com/watch?time_continue=4&v=d6MZy-2fcBw&feature=emb_logo
Posted by: atomician | Apr 2 2020 5:25 utc | 131
Hendrick Streeck is the director of the institute for virology at the university of Bonn, Germany. At the moment, he researches the spreading of the virus in the town of Heinsberg, which is the German hot spot. Here’s what he has to say: (translation from German)
Bonn / Hamburg –
The Corona crisis hits the global economy with great violence: In Germany, too, restaurants and companies have to pause for weeks, tourism stands still, nothing works in public life anymore.
The number of advertisements for short-time work has skyrocketed to an unprecedented level, and the number of unemployed is also increasing: The Federal Employment Agency expects an increase of up to 200,000 unemployed in April.
And despite the government’s aid measures, one thing is certain: the German economy will not be the same for the foreseeable future once the crisis is over. The existence of many citizens is under threat.
Hardly anyone had questioned these tough government measures, as it is about saving lives. But on Tuesday evening a well-known virologist for the first time openly raised doubts about the need for the shutdown at “Markus Lanz” (ZDF). Did our entrepreneurs have to shut down unnecessarily?
The virologist Hendrik Streeck from the University Hospital Bonn is currently carrying out a unique examination in the district of Heinsberg – the epicenter of the coronavirus. There, the expert collects both the number of infected people and the infection routes in a representative sample. The study is intended to provide answers to questions such as where the greatest sources of danger are. How exactly the virus is transmitted. How high the unreported number of infected people is. The research group around Streeck wants to publish the first results as early as next week.
The virologists had not succeeded in breeding Sars-Cov-2 in initial tests after swabbing various objects in apartments of highly infectious residents, sinks, doorknobs, but also pets such as cats. “For me it looks like the first results that a door handle can only be infectious if someone has coughed in the hand beforehand and then grabs the handle immediately.” This suggests that there is no smear infection. Keeping a distance and washing hands is therefore a very effective tool.
However, the risk of infecting someone else while shopping is considered to be low. “We see how the infections took place. That was not in the supermarket or in the restaurant or at the butcher’s. That was at the parties at the après ski in Ischgl, in the Berlin club, trumpet ’, at the carnival in Gangelt and at the exuberant football games in Bergamo.
In the current discussion about the “shutdown” and the “exit” strategies, which lead again from a standstill, such reliable facts are important. So that public life doesn’t stand still for too long.
“We talk a lot about speculation and model calculations. With these, however, only one factor has to be wrong and the whole thing collapses like a house of cards. “That is why facts are so important to make effective decisions. He was therefore surprised that the Robert Koch Institute, as the highest federal authority for infectious diseases, had not previously carried out such an investigation. He sees such tests as a duty for virologists “to find answers for the citizens.”
Did the shutdown come too quickly?
Streeck looks back at the various measures taken by the federal government, which have gradually restricted life: Larger events have been canceled, schools have been closed down to exit restrictions. “But I had already said in advance: We want to wait and see what happens. The virus doesn’t obey any politician. ”
Measures that are now decided would only be visible in the statistics in two weeks at the earliest. “You have to give this virus time so that we can see and classify the results of the measures in the long term.”
He had never heard of infections in hairdressing salons, said Streeck. But now they are closed. It is the same with supermarkets or the like. “We just don’t know that infections have taken place there. I think it’s important that we focus on what we really know – and what we don’t. ”You have to find the nuances of when exactly an infection occurs. And this must also be the guideline for reducing certain measures.
A very good way to contain the virus effectively: do a lot of tests like South Korea did. “If they tested people positively and found a cluster, then they contained the area there,” says Streeck. A nationwide curfew was not necessary there. “In my eyes, this is a very good strategy and also a strategy that is feasible in Germany. Because we have the options. ”
The virus is really dangerous for the risk groups, so “when it comes to the hospital, nursing home and old people’s home,” said the doctor. It is therefore very important to effectively protect particularly vulnerable people, with weekly corona tests for medical and nursing staff, for example. Such pool procedures are already used in transfusion medicine to test blood. So you are not new.
“It is therefore important to develop exactly such ideas. However, many experts are involved in this development, and not just individual ones. ”It is a shame that the government approached the crisis“ rather monothematically ”. Unfortunately, there is no round table with a large number of virologists, in which China is also involved.
Streeck criticizes the lack of objectives in the fight against Corona
“I see what such a curfew does to people,” explains the virologist. He himself has friends who wonder if they still have a job after the crisis. “In relation to other epidemics and viruses, I find these restrictions to be very drastic.” Before taking such measures, Streeck would have liked to think carefully: “Where do we actually want to go?” He would lack the precise definition of the objective.
“Our limit is the capacity of the hospitals. Not the number of people infected. But we never heard where our guideline was. What is our goal? Are 1000 infections a day too much? Or 100? We have to listen to the intensive care physicians who tell us where their limits are. ”They could best assess which measures are the right ones.
Marcel Fratzscher: “A good health system needs a functioning economy”
Streeck therefore supports the fastest possible discussion about an exit strategy. Marcel Fratzscher, President of the German Institute for Economic Research, explains how great the danger for the economy is at “Lanz”. He speaks of a “catastrophe” with a “rat tail of problems”. Small businesses and the self-employed could only last a few weeks despite government aid.
Anyone who receives a salary of 60 or 70 percent in short-time work can hardly stay afloat in the long term. At the same time, the economist feels uncomfortable weighing human lives against the financial damage – as many in the discussion about an exit strategy do. “Because a good health system also needs a functioning economy.”
One should not play both sides against each other, but rather find a solution that is acceptable to everyone. After six to eight weeks, the loss caused by the shutdown would become critical. And that must be avoided.
Posted by: mk | Apr 2 2020 8:15 utc | 145
|
