[lysp]

Found this report circulating on twitter from a junior doctor in Australia.

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:

https://twitter.com/char_durand/status/1221997021663387649

[kavalg]

From what I've read so far, it seems that hosts can infect healthy people 1-2 weeks before first symptoms. Something missing though is how long after the infection does the host become active in spreading the virus.

[IAmGraydon]

China has stated that they think it can be contagious before symptoms. The US has stated as of yesterday (Jan 27) that there is no evidence for that yet.

[jhrmnn]

The German case confirms it is. A woman (from Shanghai, infected by her parents from Wuhan) gave a training seminar in Munich. She transmitted the virus to her German colleage (33 yr), but started to feel ill only on her flight back to China. (The man developed symptoms over the weekend, but his state improved, so he actually went to work on Monday, was hospitalized on Tuesday.)

[MR4D]

In case anyone wants a link to a source for the above comment, here you go:

https://www.dw.com/en/germany-confirms-human-transmission-of...

[Seenso]

> China has stated that they think it can be contagious before symptoms. The US has stated as of yesterday (Jan 27) that there is no evidence for that yet.

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.

[Symmetry]

Yeah, that seems to be why the estimated R0 is so large.

[yakshaving_jgt]

Western news media are currently reporting ~100 dead and ~4,500 infected (though that number is surely much higher in reality). This is a ~2% mortality rate.

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.

[jerf]

You can naively extrapolate that, but it will be, well, a naive extrapolation. Not necessarily a bad thing, but it won't necessarily be accurate. If the virus does get into, say, the US, but it happens to only infect 20-40 year-olds through office transmission, it probably won't even be that fatal. If it happens to get into a senior home, it could be a great deal more deadly.

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.

[bildung]

I'm not qualified to have an opinion on the the virus, but the numbers could exhibit a selection bias: People with relatively mild symptoms and no death probably could be misdiagnosed as having a flu or proper influenza, which would imply that the mortality rate would actually be lower.

[sliken]

See: https://www.thelancet.com/journals/lancet/article/PIIS0140-6...

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

So among those that were able to get to a hospital 15% died.

Question is what percentage of people exposed get bad enough to want to go to the hospital?

[infinity0]

What about a 99.99% reduction in travel? Wuhan has 11 million and 1% of that is 100k, I'd imagine the quarantine is more successful than preventing 100k people from travelling.

[lsllc]

According to published reports (see [1]), 5 million people left Wuhan before the quarantine was implemented.

[1] https://www.scmp.com/news/china/society/article/3047720/chin...

[infinity0]

As I said in a different comment, 5 million people leave Wuhan every year around this time, and the model that made this 99%/25% prediction is using data that already takes into account this mass-migration.

[99052882514569]

I think their point is that with such a high reproduction number, the impact of anyone slipping through is amplified so much that reduction in travel alone will not do much to stop the spread.

[infinity0]

> reduction in travel alone will not do much to stop the spread.

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.

[jiofih]

That would have been nice. The real figure is more like -50% effective: https://nypost.com/2020/01/27/half-of-wuhans-population-fled...

[infinity0]

It's not. The model used historical figures from Jan 2017 as its input data, so would also be taking into account these New Year's mass-migrations. Also, this year's stopped at a relative early stage.

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.

[jiofih]

Are you replying to the wrong comment? The article I posted is about five million having left the city _before lockdown_, which completely invalidates any model.

[brigandish]

Figure 4 in the images from this tweet[1] (taken from the same paper above, page 8) show the effect of a 99% reduction in travel, over 65% of people in cities across China will become infected.

[1] https://twitter.com/DrEricDing/status/1220919589623803905

[infinity0]

The guy needs to Chill T F O. R0 of measles is like 12-18.

And with the city-wide quarantine people should be modelling 99.99% reductions in travel not 99%.

[brigandish]

> Dr. Eric Feigl-Ding (Eric Ding) is a health economist, epidemiologist, and nutrition scientist at the Harvard Chan School of Public Health, and an expert advisor to the World Health Organization.

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.

[1] https://scholar.harvard.edu/ericding/home

[infinity0]

Literally thousands of other people have the same qualifications as him and are also not freaking out over Twitter. I'd say the data leans more heavily in the other direction.

[brigandish]

> As a Web of Science Highly Cited Researcher, He was ranked in 2018 as among the Top 1% of all scientists worldwide.

Let's cut that down to hundreds.