Coronavirus Real Time Map
gisanddata.maps.arcgis.com
gisanddata.maps.arcgis.com
https://www.medrxiv.org/content/10.1101/2020.01.23.20018549v...
Novel coronavirus 2019-nCoV: early estimation of epidemiological parameters and epidemic predictions
> Key findings:
> ● We estimate the basic reproduction number of the infection (𝑅𝑅0) to be significantly greater than one. We estimate it to be between 3.6 and 4.0, indicating that 72-75% of transmissions must be prevented by control measures for infections to stop increasing.
> ● We estimate that only 5.1% (95%CI, 4.8–5.5) of infections in Wuhan are identified, indicating a large number of infections in the community, and also reflecting the difficulty in detecting cases of this new disease. Surveillance for this novel pathogen has been launched very quickly by public health authorities in China, allowing for rapid assessment of the speed of increase of cases in Wuhan and other areas.
> ● If no change in control or transmission happens, then we expect further outbreaks to occur in other Chinese cities, and that infections will continue to be exported to international destinations at an increasing rate. In 14 days’ time (4 February 2020), our model predicts the number of infected people in Wuhan to be greater than 190 thousand (prediction interval, 132,751 to 273,649). We predict the cities with the largest outbreaks elsewhere in China to be Shanghai, Beijing, Guangzhou, Chongqing and Chengdu. We also predict that by 4 Feb 2020, the countries or special administrative regions at greatest risk of importing infections through air travel are Thailand, Japan, Taiwan, Hong Kong, and South Korea.
> ● Our model suggests that travel restrictions from and to Wuhan city are unlikely to be effective in halting transmission across China; with a 99% effective reduction in travel, the size of the epidemic outside of Wuhan may only be reduced by 24.9% on 4 February.
> ● There are important caveats to the reliability of our model predictions, based on the assumptions underpinning the model as well as the data used to fit the model. These should be considered when interpreting our findings.
Source:
This is exactly what you can't conclude from the analysis that was done, the numbers matter significantly.
If a 99% restriction locally gives a 25% reduction globally, that tells us very little about what happens with a 99.99% restriction, which is probably closer to what has been achieved.
Sure, if you want to get precise, you can model a 0% reduction in traffic for the first 10 days, then a subsequent reduction to 99.99%. The numbers will be completely different than modelling a reduction to 99% for the whole period.
Quoting numerical estimates without understanding how the underlying model compares to reality, is just stupid.
[1] https://twitter.com/DrEricDing/status/1220919589623803905
And with the city-wide quarantine people should be modelling 99.99% reductions in travel not 99%.
I'm not usually one for relying heavily on authority, but this guy[1] is likely to know when the R0 number looks very, very bad.
Let's cut that down to hundreds.
[1] https://www.scmp.com/news/china/society/article/3047720/chin...
IIRC, the US hasn't seen the evidence for that yet, but they're also complaining that China's not sharing data with them. The CDC is getting its info from China from press briefings like everyone else.
https://www.dw.com/en/germany-confirms-human-transmission-of...
Can we naïvely extrapolate that we expect ~4,000 casualties from ~190,000 infections?
I'm not good at understanding numbers, but I'm sure someone here can chime in with a better way to read this.
And that's before we consider the spoiler of mutations, which is the real problem. In the long term, it is evolutionarily advantageous for the virus to become less lethal and eventually fade into the background as just another cold, if it isn't wiped out by aggressive quarantining. However, in the short term, many of the same things that will make it more transmissible, such as more effectively converting host systems into virus factories, or contrariwise, being more effectively hidden while still being contagious, will also make it more dangerous to the host and/or society.
But for all that, there is a sense in which the "naive extrapolation" is also the best thing we have right now based on available data. It is, at least, data-driven.
From the above URL, among 41 people who went to the hospital:
* All 41 patients had pneumonia with abnormal findings on chest CT
* acute respiratory distress syndrome (12 [29%])
* RNAaemia (six [15%])
* acute cardiac injury (five [12%])
* secondary infection (four [10%])
* 13 (32%) patients were admitted to an ICU
* six (15%) died
Question is what percentage of people exposed get bad enough to want to go to the hospital?
In fact, multiple specialist academics said ~3 days ago that only 5% of cases are confirmed, now may be less due to reports of equipment shortages, the additional time the virus has had to disperse geographically, and Chinese New Year. Therefore, we can multiply any official figures by 20x. Official figures are nearing 5,000 cases, which extrapolating from those estimates means we're at around 100,000 infections as of today.
Source: I wrote and am maintaining an animated map of the domestic spread (the one on Wikipedia), over here: https://github.com/globalcitizen/2019-wuhan-coronavirus-data... .. powered by two scrapers, main one is from DXY, which is ahead of this 'real time' map by many hours, judging by total figures.
Even if they do, why is 20 a "low" number? They are talking about estimating the current real figure, from the current laboratory-confirmed figure. 20 seems pretty reasonable if a little high, assuming the epidemic is indeed still in the exponential phase.
The same site was handling fire information for the Australian bushfires and had the same text - although that was more "realtime" as the data source was being updated much more frequently.
Fatal Errors
Unable to load https://gisanddata.maps.arcgis.com/sharing/rest/content/item... status: 502
d@https://js.arcgis.com/3.31/init.js:112:340 d@https://js.arcgis.com/3.31/init.js:140:425 f@https://js.arcgis.com/3.31/init.js:145:35
Maybe the load is a bit too heavy for the app at the moment
EDIT: after a few attempts I was able to load the map, it is nicely done!
[0] - https://twitter.com/Terrence_STR/status/1221100970521829377
1. example from another comment below: https://news.ycombinator.com/item?id=22168981
I died laughing at this! It of course is very serious and not funny, but I still laughed at this detail.
That said with a mortality rate of 2%, affecting a country like China with 1.4 billion people, even if half the people get infected, that's 14 million dead.
So overall serious on a large scale.
34157 dead from 35520883 infected, let's see how much hire a 2% corona virus mortality rate is above the normal flu mortality rate in the US:
>>> .02 / (34157 / 35520883.) 20.798596480955588
About 21 times higher.
The above comment saying that it "mostly" impacts elderly is a bit vague. The fact that within the first 100 deaths were multiple young healthy people means it is dangerous for the entire population. People in their 30s and 40s are not "old".
Citation?
source: https://www.cdc.gov/nchs/fastats/flu.htm
SARS death rate is around 10%
source: https://en.wikipedia.org/wiki/Severe_acute_respiratory_syndr...
Here's a better set of comparisons:
Deaths per 100,000:
Influenza: 2
SARS: 0.22
https://www.cdc.gov/nchs/fastats/flu.htm, https://en.wikipedia.org/wiki/Severe_acute_respiratory_syndr..., https://en.wikipedia.org/wiki/South_China
% of deaths of hospitalized people:
Influenza: ~10%
SARS: ???
https://www.cdc.gov/flu/about/burden/index.html
% of deaths of diagnosed people (~CFR):
SARS: ~10%
Influenza: 0.1%-10% per strain
https://en.wikipedia.org/wiki/Severe_acute_respiratory_syndr..., https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3809029/
The SARS mortality rate for China in 2003 was ~0.02 (349 deaths, 1.3billion population). I was being nice to you and using just the population of southern china where the sars outbreak primarily occurred.
Whether that's the most appropriate measure for this comparison isn't clear, which is why included the other measures.
Also, that's looking at the US mortality rate for influenza which is particularly low. The worldwide average is 5.9, with regions ranging from 4.5-6.2 (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6815659/).
Also, you should be more precise in your wording. 3-4 orders of magnitude worse than the flu is absolutely not true. The mortality rate for the flu is around 14.3 per 100,000. With this as a baseline, 4 orders of magnitude is 143%, which isn't actually possible unless you are aware of people being able to die more than once. Even 3 orders of magnitude is 14.3%, which is higher than SARS. The current estimate is around 2%, which is more like 2 orders of magnitude.
And I'm not implying that people 47 and older arent worth caring about. But that group is always tends to have more severe health problems, which is why people should take those numbers with a grain of salt.
They've reported that strong symptoms occur in only 25% of cases, so if you get it chances are you're just experiencing having a cold (which is caused by a coronavirus).
You say "even" as if it's a conservative estimate, but to me 50% of the country's population being infected by the same virus sounds extreme. Wouldn't this be on the extreme side?
If we assume that this virus is as infectious as the common cold, we can expect the infection rate within a household to be approximately 25% of contacts. That means that of the people you LIVE with, 25% will catch it from you.
50% infection rate in a large, distributed population like China would be very extreme.
https://www.ijidonline.com/article/S1201-9712(19)30354-6/pdf
This is all well understood, and studied in considerable depth.
The media overreaction is typical, but here at HN we are better than that and we should strive to rely on established science (where available). No need to throw out extreme or unreasonable numbers!
Colds are quite different to flues.
I used influenza because it was convenient in terms of available research, I could grab in a minute or two, but I'm sure if you care to look you can find similar data for rhinovirus.
Rhinovirus and influenza have very comparable R0's. R0 is the epidemiological measure of the "contagion" factor of a pathogen.
We could try to make the comparison you are suggesting instead. But to make a valid comparison we need to consider:
Is influenza infectious for 10 days before any symptoms show?
How many people get the influenza vaccine each year? I.e. is there some herd immunity built up?
The symptoms of influenza are universally nasty for everyone infected. No one with the flu is walking around and going to work (if you think do then you don't have the influenza virus you have a cold). It seems that at least some people with this new corona virus just have cold like symptoms and therefore will not by default be self-isolating like people with influenza naturally do.
Influenza is a useful proxy for the cold, because we have plenty of data on influenza strains with R0 very close to the cold.
Perhaps the most contagious disease we've ever encountered, the measles, has an R0 around 18. Before the 1960's, when the vaccine was licensed, we saw incidence rates as high as the .8% range yearly for measles. That is 20x more than influenza, but still orders of magnitude short of the 50% number you threw out there.
The cold and influenza both range from R0=1.3, to perhaps 6 on the very high end of estimates. 50% just isn't reasonable by any measure.
As for comparisons to this nCoV, it's still very early days and there are many unknowns. Still, there is no evidence to support an R0 even remotely close to the measles. 50% simply isn't plausible or reasonable, based on everything we know about viruses and epidemics.
And not just that. I have a septum deviation, so I basically have a runny nose all winter long, even though, most of the time, no viruses are involved.
Even directly after surgery, with all the related inflammation and dried blood and gross up there, I immediately noticed a massive airflow difference.
Do it.
Summary: We know of deletion event in SARS Orf8, and amino acid change in ebola glycoprotein. In both cases, it decreased viral replication or virulence.
Early disease death rates are always overestimated because people tend to report only the worst cases of it (even more if the usual symptoms look like a cold). But there is no way to know by how much we are overestimating it.
I'm staying inside for now...
Rather, consider using a heatmap to show the proportional density of observations in a way that better highlights their geographic distribution.
The media sell stories. People click on those stories, the stories that get clicked on most get other stories written on the topic. This is how the industry works. It's actually quite a simple model.
Human beings are fucking terrible at risk assessment. What do people here think the risk is for them of dying of coronavirus compared to say dying in a traffic accident? Which is more likely to happen to you by a few magnitudes?
That said, I don't trust the numbers so at this point it's hard to say. My gut says there are way more infected people than reported, but deaths are probably less likely to be underreported at that scale. Which means mortality rate would actually be significantly lower than 2%. On flip side, the chance of getting infected would be much higher.
Tough for most to outrun/outsmart an epidemic.
Read/watch some of the Twitter whistleblowers.
Govs have no reason to tell everyone the true risks - and every reason to hide them.
That's absurd considering how many people are unwilling victims of accidents, how exactly would you outsmart it? Never leave home?
It's a thing.
https://en.wikipedia.org/wiki/Illusory_superiority
Regardless, let's turn that comment around: could it not similarly apply to hygiene and risk of infection? Just be an above-average smart citizen and you'll be fine, yes?
Doesn't sound like a lot, but an R0 of >2 is concerning because of what it means for total infected. 2% of everyone within range of access to modern transportation is an awful lot of people.
But there are two few confirmed cases to understand the mortality rate. Might be there are additional factors unique to China that is causing a higher death rate.
I was reading somewhere that the doubling in cases in China was mainly due to increasing capacity for hospitals to diagnose the infection rather than an actual measurement of the spread of the disease.
People are being turned away from hospitals. Dead patients are not being post-humously tested for coronavirus. Most patients hospitalized arent yet recovered (or dead) yet etc...
The only rate we know so far is the dead/hospitalized rate, which is around 13% - which comperable to SARS.
Great presentation tool.
Why is this coronavirus stuff not flagged also? There are currently 3 articles on the front page related to this. It just slipped or is there a reason for this?
Kobe's death, while tragic, is more along the lines of current event discussion.