Talk:Timeline of the COVID-19 pandemic in January 2020/Archive 1

Title

I know it's not an influenza virus but a corona virus. Had to come up with something -rename as appropriate. Wuhan Pneumonia, perhaps? kencf0618 (talk) 17:25, 23 January 2020 (UTC)[reply]

IMO, the best title would be Timeline of the 2019-20 coronavirus outbreak.MSG17 (talk) 18:58, 23 January 2020 (UTC)[reply]

Probably. "Wuhan Flu" is just a placeholder.kencf0618 (talk) 19:25, 23 January 2020 (UTC)[reply]

What happened to the stats of quarantined people ? of cured people ? etc.

This page was until this afternoon a very complete informative page with an excellent overview of all official stats, not only the total of confirmed cases, but much more.

Why has this been changed ? And can it please change back ? 20:59, 29 January 2020 (UTC) — Preceding unsigned comment added by PieterJansegers (talk • contribs)

Orphaned references in Timeline of the 2019–20 outbreak of novel coronavirus

I check pages listed in Category:Pages with incorrect ref formatting to try to fix reference errors. One of the things I do is look for content for orphaned references in wikilinked articles. I have found content for some of Timeline of the 2019–20 outbreak of novel coronavirus's orphans, the problem is that I found more than one version. I can't determine which (if any) is correct for this article, so I am asking for a sentient editor to look it over and copy the correct ref content into this article.

Reference named "auto":

From Hainan: 李齐. "Province burns bright as investment hot spot amid industrial boom - Chinadaily.com.cn". www.chinadaily.com.cn.
From Singapore: "Singaporeans have world's longest life expectancy at 84.8 years". Straits Times. 20 June 2019.
From 2019–20 Wuhan coronavirus outbreak: hermesauto (20 January 2020). "Wuhan virus: China announces more confirmed cases, including in Shanghai and Guangdong". The Straits Times. Retrieved 20 January 2020.
From Centers for Disease Control and Prevention: "Public Law 104–208" (PDF). Congressional Record. September 30, 1996. ... none of the funds made available for injury prevention and control at the Centers for Disease Control and Prevention may be used to advocate or promote gun control.

I apologize if any of the above are effectively identical; I am just a simple computer program, so I can't determine whether minor differences are significant or not. AnomieBOT ⚡ 23:31, 23 January 2020 (UTC)[reply]

Unnecessary Sentence about Plague Inc

Is the sentence "Ndemic Creations received so many inquires about the actually occurring epidemic that they directed users via their Twitter account to the CDC." under the date January 23, really necessary on this timeline about the Wuhan Virus outbreak? I think it should be removed. --Colin dm (talk) 03:11, 24 January 2020 (UTC)[reply]

Ndemic Creations gave a talk to the CDC a few years ago because the CDC recognized that the Plague, Inc. was not only a fairly good model but served as an unusual (and effective) platform for public health. They've done similar with the Corrupted Blood incident and the Zombie trope. Even the DOD has used zombies as an out-of-the-box testing exercise for its command structure. Click around -some games are silicon Petri dishes for, and hence to some extent of, human behavior. Click around: Inadvertent theoretical models in the wild have been quite useful for professionals. kencf0618 (talk) 11:07, 24 January 2020 (UTC)[reply]

BBC Omnibus Jan. 24th

Quite a bit to parse here, so I'll just leave this here:

https://www.bbc.com/news/world-asia-china-51230011

Present or past tense?

Should the present or the past tense be used across this article? Currently, it is mixed.--Adûnâi (talk) 13:00, 24 January 2020 (UTC)[reply]

check MOS:TENSE, should be past since encyclopedia and will be the record of current event.JuanTamad (talk) 22:59, 25 January 2020 (UTC)[reply]

2009 flu pandemic timeline: A Precursor

https://wiki.riteme.site/wiki/2009_flu_pandemic_timeline

These manifold events are very different beasts, but 46.1% by edits of the swine flu timeline is my work (full disclosure) and may well serve as a template. kencf0618 (talk) 19:17, 24 January 2020 (UTC)[reply]

We should be using UTC

Considering many of the edits on this page are concerning cases outside of China, I think we should make UTC official for this page to clear up any confusion about dates. --Colin dm (talk) 20:30, 24 January 2020 (UTC)[reply]

Since we are talking about an outbreak focused and started in China, using UTC+8 seems more logical especially for talking about how it started. I would like to add a template for that. Artoria 2e5 🌉 08:43, 25 January 2020 (UTC)[reply]

Using Chinese UTC+8 for a case on the other side of the globe does not make sense. It should be either UTC or local time. I vote for local time. This will make it more consistent with official reports and sources. Balmclothe (talk) 00:15, 26 January 2020 (UTC)[reply]

All charts on this page are outdated

All of the charts show a third the number of cases since they haven't been updated. We should remove them.

For the two topmost, there is R code. I could recreate, but the sources are in Chinese and the pages don't have the numbers as far as I can see - here for example. Would have to get data from another source unless someone can read this. TBut not sure these are the best graph types to continue with. Wouldn't something like the nytimes graph be preferable? JuanTamad (talk) 23:52, 25 January 2020 (UTC)[reply]

I like the look of the current charts, and think they look nice. I think the best idea is to just switch to a different source for the graphs, and then recreate them. Syryquil1 (talk) 02:11, 26 January 2020 (UTC)[reply]

This might help https://www.cgtn.com/special/Battling-the-novel-coronavirus-What-we-know-so-far-.html Nebakin (talk) 11:33, 26 January 2020 (UTC)[reply]

Statistics go on the main page

Just a reminder since I keep seeing this problem: statistics about the number of infections (eg. x,xxx number of cases reported today) do not belong here. They belong on the main page unless the cases have some precedence (eg. first/second/third case reported in country). --Colin dm (talk) 03:28, 26 January 2020 (UTC)[reply]

Why? Seems a timeline should include the number of deaths that day. Especially to spot a trend. I don't see 14 deaths anywhere on the main page. It is good to present information in different ways. Daniel.Cardenas (talk) 03:37, 26 January 2020 (UTC)[reply]

The timeline needs a new, flagged section which simply updates the number of cases and death per political unit. I haven't the chart-fu, though. kencf0618 (talk) 14:29, 26 January 2020 (UTC)[reply]

Missing 2 new cases for Taiwan on 2020-01-24

On 2020-01-24, Taiwan reported 2 new confirmed cases, for a total of 3, which so far are missing from the timeline. Sources:

https://www.mohw.gov.tw/cp-4636-51207-1.html (Chinese)
https://www.cdc.gov.tw/En/Bulletin/Detail/xzPuzBOQiW0Gjw1MfyNKnw?typeid=158 (English)

Could someone please edit those in? I don't feel quite comfortable editing citations yet, especially in an article where they're critical. LukasGraf (talk) 18:30, 26 January 2020 (UTC)[reply]

Stop adding statistics for the number of cases when the day has not finished

I will remove any changes that add unfinished days as that is the style on the main article. If you want to change this start a talk page on the main article. --Colin dm (talk) 04:18, 27 January 2020 (UTC)[reply]

Please use the term "exponential" when epidemiologically appropriate.

Exponentals are best plotted on semi-log graphs. Polynomial functions are best plotted on log-log graphs.

The reason this is epidemiologically important is that exponentials immediately imply human-human transmission. Whereas polynomial relations imply the transmission is from sources external to the human population. I get R^2 values of 0.993 for log(deaths) vs time.

The growth rate is exponential. — Preceding unsigned comment added by AndyFroncioni (talk • contribs) 13:25, 27 January 2020 (UTC) [reply]

Epidemic & Pandemic/PHEIC Sections

I have taken the liberty of collapsing the Epidemic section, labeled as such, as it was becoming cumbersome and shall become more so. That's the nature of the beast; it's a time line of a very complicated, volatile, global event. Lots of moving parts.

We should soon have maps soon of confirmed cases and deaths -I suggest that we place them every other day, or perhaps weekly, depending on how comprehensive we want to be. kencf0618 (talk) 16:32, 28 January 2020 (UTC)[reply]

Labeling the entire thing as pandemic phase is inappropriate since it has only recently become an epidemic. It would be better and more accurate to relabel it as "Chronology".--Colin dm (talk) 17:37, 28 January 2020 (UTC)[reply]

It's an epidemic now. If and when that status changes (which dependent on any official announcement by the WHO), we'll need a new section accordingly savvy? kencf0618 (talk) 18:10, 28 January 2020 (UTC)[reply]

What's the source for the 4,475 infected number?

It doesn't seem to be https://gisanddata.maps.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6 (numbers don't match) or http://www.nhc.gov.cn/xcs/yqtb/list_gzbd.shtml (the numbers don't match).

--190.189.120.70 (talk) 01:44, 28 January 2020 (UTC)[reply]

for up to the minute current numbers, I look at bnonews.com SailBelow (talk) 06:42, 29 January 2020 (UTC)[reply]

Formatting suboptimal

On the desktop end, a graph is shown at the left of the page and makes the whole article misaligned. I can't quite figure out how to fix that though.

Rethliopuks (talk) 08:57, 29 January 2020 (UTC)[reply]

Future projections in cases

Not sure who added the future projections to the case chart, but I am opposed to doing that. Disease progression can be very unpredictable and reporting projected future totals has a lot of potential to confuse readers. Dragons flight (talk) 09:13, 29 January 2020 (UTC)[reply]

Seems that they have already been removed. Dragons flight (talk) 09:55, 29 January 2020 (UTC)[reply]

Maps regarding the measures taken in Hong Kong / Map Taskforce

Hi!

Is there any map(s) that transmits the measures taken in Hong Kong, for example about the ports and checkpoints that are closed?
And what about other maps of other regions?
Is someone able to create them on a .svg basis?
Are there open-source maps related to travel restrictions and/or quarantines in effect?
Can we get together a map taskforce for this?

I do not have the knowledge required to create them but I am able to do some research and collect data for the map creators/MapTaskForce :)

What do you think about it?

Thank you. FranciscoMMartins (talk) 14:30, 29 January 2020 (UTC)[reply]

Fishing back provincial data !

Done Yug (talk) 21:37, 29 January 2020 (UTC)[reply]

Hoy ! I think Rethliopuks data and formating has value ! So I fished them back and store these data into templates {{2019-20 Wuhan coronavirus data/China medical cases}}, {{2019-20 Wuhan coronavirus data/China medical cases by province}}. This template will be accessible via the root template associated with that topic. Yug (talk) 21:24, 29 January 2020 (UTC)[reply]

Done, visible in the "Data and figures" row of {{2019-nCoV}}. Yug (talk) 21:37, 29 January 2020 (UTC)[reply]

Possible plot with regressions

I am considering adding this log-linear plot of confirmed cases and deaths versus time. The close linear fit for confirmed cases suggests the epidemic is growing exponentially.

Galerita (talk) 02:25, 28 January 2020 (UTC)[reply]

If you keep it updated, it's a great addition. --Eric1212 (talk) 04:14, 28 January 2020 (UTC)[reply]

Great!--Ifwest (talk) 16:59, 28 January 2020 (UTC)[reply]

What is the rationale for starting the plots with January 16th? SailBelow (talk) 04:06, 29 January 2020 (UTC)[reply]

The confirmed cases are continuous starting from 16 January and the update of the deaths from 2 to 3 (on 19 January) occurs from news reports cited in the body of the article. I also wanted to illustrate the "exponential phase", but I'm also considering the best way to present the data.

Galerita (talk) 00:41, 30 January 2020 (UTC)[reply]

Also, the plot should extend some days (about a week) into the future. On any given day, there are people who are newly infected. They are starting in their incubation phase and are unknowingly sick. After their incubation phase, they will become symptomatic and start to feel sick. Only at that point would they pursue a diagnosis and become a 'confirmed case'. Thus on any given day, there exists a week's backlog of infected patients. Extending the plot into the future by a week would show the anticipated number of new cases.SailBelow (talk) 06:40, 29 January 2020 (UTC)[reply]

Extrapolating beyond the data is dangerous. It assumes the the current rate of exponential growth will continue into the future, which is doubtful. I'm also trying to balance presentation of statistics with original research.

Galerita (talk) 00:41, 30 January 2020 (UTC)[reply]

I get semi-log coefficients of:

 log10(deaths)    = 0.18145 * x + 0.63971  (R^2 = 0.988)
 log10(confirmed) = 0.17784 * x + 1.51956  (R^2 = 0.995)

Growth is about 18% per day for deaths and confirmed cases. — Preceding unsigned comment added by AndyFroncioni (talk • contribs) 17:00, 28 January 2020 (UTC)[reply]

Andy Froncioni (talk) 17:01, 28 January 2020 (UTC)[reply]

I'm using R and my regressions are in log_e. If growth is 18% that means 1.18^4 = 1.94 i.e less than 2. So doubling time would be less than 2 days. Look at the trend in cases and deaths on https://wiki.riteme.site/wiki/Timeline_of_the_2019%E2%80%9320_Wuhan_coronavirus_outbreak#Cases_Statistics Doubling time is clearly about 2 days or less. I get similar R^2 values to you.

Galerita (talk) 01:02, 30 January 2020 (UTC)[reply]

From my experience during the SARS epidemic, the data for deaths will be more accurate and more timely. Andy Froncioni (talk) 17:17, 28 January 2020 (UTC)[reply]

I think that's basically correct, especially for a slow moving epidemic. There appear to be a shortage of testing kits, so confirmed cases are affected by how fast suspected cases can be tested. Provided deaths are correctly attributed to coronavirus they are a better indicator of the long-term trend of the epidemic, but they lag cases by approximately 10 days, which appears to the average time between becoming a confirmed case and dying, i.e the average survival from diagnosis. Also it is likely that some earlier deaths were attributed to pneumonia. So on balance "confirmed cases" are a better leading indicator of the growth of the epidemic at the moment. I put both on the plot to give the best compromise.

Galerita (talk) 00:55, 30 January 2020 (UTC)[reply]

Cases Statistics

I really appreciate this page and check it daily. I think there is a typo on the table of confirmed cases for Jan. 29--it says 7771 but the graphic above it shows 7711 so I checked the Chinese reference and it says 7711 there. I am not an editor and don't want to mess anything up so I'm just letting the editors know. Jamgypsy (talk) 01:33, 30 January 2020 (UTC)[reply]

A broken reference

Reference [34] is broken, it has ref name=12 but no other reference uses it.--Adûnâi (talk) 06:48, 30 January 2020 (UTC)[reply]

Flags on links

Flags are not supposed to be added to links to countries outside of tables. It isn't standard practice on Wikipedia and adds nothing of value to the article.--Colin dm (talk) 21:59, 30 January 2020 (UTC)[reply]

Table Captions?

The table captions can be improved by defining what is meant by the column names. Released being released from quarantine? Or released from the hospital? Recover reporting needs a reference and definition. Plabs (talk) 22:00, 30 January 2020 (UTC)[reply]

Mainland China section

I have just split up the original chronology section to create a dedicated section for mainland China. This is similar to how the section for outside MC got created. Many issues still need to be resolved re: the two sections.

My idea of what content belongs to which sections is roughly follows:
1. Chronology: any event belonging to, or not itself belonging to but significantly affecting, the disease itself.
2. inside MC: any event related to the outbreak but in nature not an event caused by the disease itself but human actors in mainland China.
I have left some material in both sections because I couldn't decide how best to organise it. Please also rearrange any material as you see fit.

The split is beneficial for clarity of reading and information input and organisation, although some splitting of the medical situation from the human response might mean the reader needs to cross-reference (more now; not just for oMC events) within this article. It is now clearer as well what content has been covered by the article. As one example, in the former format, the 28 January subsection has several paragraphs of content. Now it appears that this article does not contain any MC responses on that date.

Rethliopuks (talk) 14:11, 31 January 2020 (UTC)[reply]

One section for the days events, with subsections?

How do you feel about the two different sections: Outbreak Chronology and Reactions Chronology? Would it be better one section per day and sub section for each? That way only have to look at one section for the days events. What do think? Daniel.Cardenas (talk) 00:30, 31 January 2020 (UTC)[reply]

I'd second this. The current organisation is going to get progressively more awkward as days pass. However, I wonder if it is at all possible to hide sections so that if you want to, you can focus on a complete chronology of just one of the parts. Rethliopuks (talk) 14:19, 31 January 2020 (UTC)[reply]

Separate Months

The months should be delineated in some way. kencf0618 (talk) 17:39, 1 February 2020 (UTC)[reply]

Very recent news sources/content from Xinhuanews.com

Hello!

I send you some links for some news posted recently on Xinhuanews.com [English]. Is anyone interested in using them or a couple as sources for new content on the article?

Is anyone interested in using them or a couple of them as sources for new content on the article?

Can you suggest/send us more sources on other news agencies from the P.R.C./Taiwan?

Chinese premier stresses ensuring medical supplies in fighting epidemic

"During video calls to medical supplies manufacturing enterprises, Li asked about their production and extended greetings to the employees who work during the holiday. He also stressed the importance of ensuring the supply of daily necessities for the public."

China to increase financial support for anti-virus campaign

"Pan said that the PBOC will provide relending funds of 300 billion yuan to national banks and local banks in the worst-hit regions, which will then grant credit support at favorable interest rates to key manufacturers of medical supplies and daily necessities."

"Source: Xinhua| 2020-01-31 21:02:21|Editor: Xiaoxia|"In face of the outbreak of pneumonia caused by the novel coronavirus, Anhui Province has organized a psychological aid medical team and started a 24-hour hot line to help those affected by the epidemic. (Xinhua/Huang Bohan)"

Chinese prosecutors urged to toughen crackdown on epidemic-related crimes

"Crimes including fabricating coronavirus-related information that may lead to panic among the public, making up and spreading rumors about the virus, sabotaging the implementation of the law and endangering public security will be strictly cracked down upon to ensure the epidemic prevention and control work is conducted in an orderly way, the circular said. It also stressed harshly punishing illegal hunting of wildlife under state protection, as well as improving inspection and quarantine measures for fresh food and meat products. "

Chinese epidemiologist identifies ways to treat critically ill coronavirus patients

"China has found effective ways of treating critically ill patients who are affected by the novel coronavirus (2019-nCoV), renowned Chinese epidemiologist Li Lanjuan told Xinhua. [...] The "two balances" refer to maintaining water electrolytes, acid-base balance and micro-ecological balance, Li said. "The treatment of critically ill patients is of great significance. Only by lowering the mortality rate can we allay the fears of the public," she said.""

China Focus: More provinces extend business, school suspension for coronavirus control

"A growing number of Chinese provinces and cities including Heilongjiang, Shandong, Guizhou, Hebei, Hunan, Shanghai and Xi'an have taken similar measures by asking enterprises, except those involved in key sectors, not to resume their work earlier than at midnight on Feb. 9, in a move to reduce crowds and prevent the spread of the coronavirus outbreak."

"In the northwestern Chinese city of Xi'an, local authorities have asked enterprises to arrange their production flexibly, while employees were encouraged to work at home."

"According to the provincial government of central China's Hunan on Friday, primary and secondary schools and kindergartens shall not start their spring semesters earlier than Feb. 17, and colleges and universities shall not resume class earlier than Feb. 24."

"Local authorities also said that all kinds of schools (including training institutions) in Hunan are not allowed to organize any form of offline teaching, training or collective activities."

::Workers make protective suits to help combat novel coronavirus

"Source: Xinhua| 2020-02-01 21:25:28|Editor: Xiaoxia"

"Workers make protective suits at a workshop of a company in Harbin, northeast China's Heilongjiang Province, Feb. 1, 2020. Resuming production on Jan. 26, the company's first batch of 400 protective suits has passed through acceptance check and will be put into use. (Photo by Zhang Tao/Xinhua)"

"Noting that World Health Organization (WHO) Director-General Tedros Adhanom Ghebreyesus has spoken highly of China's strong measures against the epidemic and expressed confidence in China's victory, Wang pointed out that the WHO believes that the Chinese actions have gone well beyond the requirements by the International Health Regulations and recommendations by the WHO, and that the WHO disfavors imposing travel or trade restrictions on China."

"Wang expressed his belief that New Zealand will fully respect WHO's important recommendations, maintain normal exchanges and economic and trade cooperation between the two countries, and firmly support China in preventing and controlling the epidemic."

"Peters said he understands and respects China's position and believes that it is necessary to maintain normal exchanges and personnel flows, adding that New Zealand will continue to communicate and coordinate with China to properly respond to the challenge of the epidemic."

"The three drugs are Remdesivir, Chloroquine and Ritonavir. They are now under relevant procedures to gain approval for clinical use, said Hubei Daily on Wednesday."

Any idea on how to better format this list, please?

FranciscoMMartins (talk) 23:23, 1 February 2020 (UTC)[reply]

WHO Situation Reports

https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports/

The deadline is 10:00 a.m.; Geneva is UTC+1 (CET), and UTC+2 (CEST) for Daylight Saving Time.

Have fun reconciling the figures.

kencf0618 (talk) 03:12, 2 February 2020 (UTC)[reply]

A Matter of Usage

I understand "overtake" to mean "to catch up to, as in a race," and "take over" to mean "take control of." Thus, in the clause reading, "Wuhan's Epidemic Prevention and Control Headquarters[450] arranged for Jointown Pharmaceutical to overtake the warehouse management of Wuhan Red Cross...," it seems the intention is to say, "Wuhan's Epidemic Prevention and Control Headquarters[450] arranged for Jointown Pharmaceutical to take over the warehouse management of Wuhan Red Cross...." I think the page is protected, so I leave it to someone else to decide whether the text should be changed.

Thank you, to everyone involved in maintaining this very useful page.

--2600:1700:10:A020:3876:4727:8E4C:2CEF (talk) 00:31, 3 February 2020 (UTC)[reply]

"New confirmed cases of coronavirus in mainland China by provincial divisions" numbers are wrong

Hey, these numbers are definitely wrong, not only from the third of February but also from the second of February. The ones with the second of February don't match the ones from the WHO and the ones from the third of February can also not be true, according to this official report by the Chinese government, we have new 890 cases outside of Hubei: http://www.nhc.gov.cn/xcs/yqfkdt/202002/bfebf84fb88248e4a4eba61295e5882b.shtml

I guess the Hong Kong, Taiwan and Macau numbers are included too but even if you substract them you shouldnt't have less than 880. Would be very nice if only the numbers of the situation report of the WHO are used, or none at all, since it can be very confusing and dangerous. It suggests that the numbers outside of Hubei are decreasing while they are not. Thanks. — Preceding unsigned comment added by 130.83.136.149 (talk) 11:06, 4 February 2020 (UTC)[reply]

Chinese citations

I'm seeing a lot of Chinese citations here, which is great unless you don't know Chinese. Someone bilingual should pair each one with a corresponding English citation (which granted would be quite a job). Curiously, I see that this article's language box has to date Czech, Persian, Portuguese, Thai, Turkish, and Vietnamese, but not Chinese. kencf0618 (talk) 17:06, 5 February 2020 (UTC)[reply]

新型冠状病毒（2019-nCoV）according to WHO. kencf0618 (talk) 18:24, 5 February 2020 (UTC)[reply]

Better section hierarchy needed

As the days pass, the table of contents keeps on growing. I think we should consider adding additional section headers for each month (December, January, etc.), then keep subsections as they currently are. We can then use {{TOC limit}} to hide all the specific days. Bobbychan193 (talk) 09:11, 2 February 2020 (UTC)[reply]

I agree with this 100%, it's getting unwieldy to the point of absurdity. Bobbychan193 (talk), are you going to go ahead and execute this? Andrew Englehart (talk) 22:57, 3 February 2020 (UTC)[reply]

You or anyone can go ahead and do it. Bobbychan193 (talk) 23:04, 3 February 2020 (UTC)[reply]

I've placed the collapse top & bottom templates to bracket the appropriate sections. kencf0618 (talk)

@Kencf0618: The decision to conceal article content needs to be reversed, in accordance with MOS:COLLAPSE. @Bobbychan193: If the article is getting unwieldy, then content should be split out to separate articles, per WP:TOOLONG. EP111 (talk) 03:05, 5 February 2020 (UTC)[reply]

Yes, I've come across this protocol before. Offhand I don't know which is the largest timeline on Wikipedia, but there needs to be one-stop shopping for this timeline (which is bound to become one of the lengthier ones). Arguably it's the exception to the rule, but I'll let consensus figure it out. kencf0618 (talk) 15:42, 5 February 2020 (UTC)[reply]

@Kencf0618: I'd suggest just splitting out the reactions sections, to a separate article, as an initial measure. The outbreak chronology should stay as one article, for the time being. When the chronology gets long enough, it'll probably be necessary to split it into separate three-month-long articles, with this article being a collection of {{main}} templates. EP111 (talk) 10:12, 6 February 2020 (UTC)[reply]

Airlines, Border Closings, and Universities

I haven't seen the numbers on any of these. A lot of heavy institutions are responding to the Wuhan coronavirus (there were no border closings or massive disruption to travel or trade during the Swine Flu pandemic). kencf0618 (talk) 15:19, 6 February 2020 (UTC)[reply]

Extrapolation

Regarding projections about the future course of the outbreak, please consider the following curve which uses only data from the "confirmed cases" table:

c=14380\left(1+\operatorname {erf} \left(\left(t-17\right)\cdot {\frac {45\cdot 1+62\cdot 2+121\cdot 3}{4000}}\right)\right)

where t represents days since 15 January 2020. You will find c agrees well with the data. Additionally, this formula predicts no new cases after 23 February. Civilization will not end: very good news! — Preceding unsigned comment added by 184.155.206.191 (talk) 23:49, 3 February 2020 (UTC)[reply]

Original research. kencf0618 (talk) 00:09, 4 February 2020 (UTC)[reply]

@Kencf0618: If your formula is correct and there are no new cases after Feb 23, have you calculated how many people would have been infected by then? And assuming that for example by Mar 15 all confirmed cases have been resolved (died or recovered), have you calculated the final death toll? Can you calculate it? --AAAAA (talk) 15:46, 4 February 2020 (UTC)[reply]

Not my doing. kencf0618 (talk)

The fit seems to be diverging from the number of current suspected cases, it is already 15% off 1 or 2 data points down the line, so it can probably be used for interpolation but is not necessarily that good at predicting the number of suspected cases in the future. If the current growth continues, we will know for sure in 2-3 days if the number of suspected cases surpasses the predicted 30000 or not.178.83.152.198 (talk) 11:31, 5 February 2020 (UTC)[reply]

@178.83.152.198: But isn't the formula about the confirmed cases? If daily growth stays at 15% there will be 500,000 infected before the end of February--AAAAA (talk) 23:41, 6 February 2020 (UTC)[reply]

2019-02-07 Proposal for reformating the page

simplified Table of Contents:

(NEW SECTION) Geographical timeline maps
Time-histogram of cases and fatalities (orange and red histogram)
(NEW SECTION)Semi-log plots of Daily and Cummulative Confirmed Cases in Mainland China
Time-Table of Mainland China Cases (currently in TOC as Case Statistics)
Highlights of Outbreak Chronology (move sub-TOC to that section)
Events, reactions, and measures in mainland China (again, move sub-TOC to that section)
Reactions and measures outside Mainland China (again, move sub-TOC to that section)
(possible new section) Media Coverage

remove the image of the virus, it is irrelevant to the timeline theme of this page.

or, recaption it as an artistic image. The two people credited on CDC with the image are medical illustrators.
or, if you know it is data based, state the source/type of data that is rendered and include "false-coloring" in caption

move the (excellent) map "Date when first case..." to a new section

move the (excellent) semi-log plots to a new section

remove the global map "Dark red denotes country)

or, explain the timeline relevance/nature of the data

remove the media "Health Commission: 440 cases" (occurs twice on the page); "Infectious Disease Expert ..."; "Hong Kong confirms ..."

or, move them to a new section
or, explain their special importance

move the (excellent) provincial-growth map (the animated map) to be alongside the first-case map

--SailBelow (talk) 21:56, 7 February 2020 (UTC)[reply]

I suggest splitting the article where possible or reduce the content while reformatting the page as it has pointed out in Wikipedia:Teahouse#What_happened_that_the_refs_just_vanished_and_got_replaced_by_template_links? that the page is currently in Category:Pages_where_template_include_size_is_exceeded, which results in the Reflist and the Navbox templates not being shown properly. robertsky (talk) 02:38, 8 February 2020 (UTC)[reply]

Update, I have broken up the article by month to get around techinical limitations. robertsky (talk) 03:49, 8 February 2020 (UTC)[reply]

Many thanks. kencf0618 (talk)

New error in case statistics table.

Case fatality ratio (CFR) is not equal to D/(D+R), unless the outbreak is finished (no new cases). CFR is well defined and used in the medical/scientific community as (Number of deaths)/(Number of CONFIRMED cases). CFR definition is in general not to be modified to include potential or suspected cases. Unsurprisingly, D/(D+R) is not commonly used in the scientific community (I have failed to find one scientific article that uses it) e.g. 2019-nCoV is neither 63.9% nor 51.6 nor 40.2% lethal and neither is its lethality changing that much over a week ! In contrast, CFR has been from 2.35% down to 2.08% during the past week. CFR should at least have a column of its own. — Preceding unsigned comment added by TheRightKindOfDoctor (talk • contribs) 14:52, 4 February 2020 (UTC)[reply]

@TheRightKindOfDoctor: I agree that CFR is not equal to D/(D+R) unless the outbreak is finished. AND ALSO every single case has to be resolved. Even if there are no new cases, the final and correct CFR cannot be calculated just when there are no new cases, as there's some time in which the current confirmed cases have not been resolved. Also, it seems that D/(D=R) is not commonly used in the scientific community, and I find that incredible. More incredible is to tell the public that the CFR is only 2% when in reality it seems to be 10% right now. If you have a rate of growth of confirmed cases of 15% or more and 15 days average to "resolve" a case, you are talking about a group of unresolved confirmed cases of 7X (1.15^15-1) the quantity of resolved cases. So the deaths/confirmed ratio will obviously be much lower than the real CFR ratio. I honestly never analyzed these numbers before, and now that I did I cannot begin to understand how bad can the "epidemiologists" and the "scientific community" be. I understand the politicians, as they always lie to "keep calm". But scientists should know better! The D/(D+R) might not be "commonly used" but I think in a pandemic with this parameters, it is useful. It will keep coming down and eventually will approach the real CFR, even if the confirmed growth rate keeps being high. The Deaths/Confirmed ratio will keep being artificially depressed as long as the daily growth is high and only show the true figure when the growth tapers off substantially. I think that's VERY BAD!--AAAAA (talk) 00:03, 7 February 2020 (UTC)[reply]

Should we remove the numbers though? I feel like whether it's CFR or D/(D+R), it's not exactly relevant information at this point in time. Orangewarning (talk) 11:42, 5 February 2020 (UTC)[reply]

See above discussion. The column should be removed.―― Phoenix7777 (talk) 11:51, 5 February 2020 (UTC)[reply]

@Orangewarning: CFR is a standard well defined indicator in epidemiology/medicine i.e. "Case fatality rate measures the severity of the disease that causes death" (typical use and definition https://www.sciencedirect.com/topics/medicine-and-dentistry/case-fatality-rate and the very nice didactic definition from https://www.britannica.com/science/case-fatality-rate). CFR is different from mortality rate, attack rate, etc and it assumes nothing about the onset/delay until death or survival/recovery. Bottom line: D/(D+R) should DEFINITELY be removed and MAYBE the CFR indicator could be added although (1) it is very easy to find elsewhere, (2) its relevance is arguable this early in the outbreak ... as you pointed out.TheRightKindOfDoctor (talk) 13:47, 8 February 2020 (UTC)[reply]

@Orangewarning: :@Phoenix7777: :@TheRightKindOfDoctor: I agree that CFR is not equal to D/(D+R) unless the outbreak is finished. AND ALSO every single case has to be resolved. Even if there are no new cases, the final and correct CFR cannot be calculated just when there are no new cases, as there's some time in which the current confirmed cases have not been resolved. Also, it seems that D/(D=R) is not commonly used in the scientific community, and I find that incredible. More incredible is to tell the public that the CFR is only 2% when in reality it seems to be 10% right now. Imagine you get infected and you get told that only 2% die... you might be worried but think that only the old and frail people will die. But then every 1 of 10 people actually die. Surprise! I am sure you wouldn't like that. If you have a rate of growth of confirmed cases of 15% or more and 15 days average to "resolve" a case, you are talking about a group of unresolved confirmed cases of 7X (1.15^15-1) the quantity of resolved cases. So the deaths/confirmed ratio will obviously be much lower than the real CFR ratio. I honestly never analyzed these numbers before, and now that I did I cannot begin to understand how bad can the "epidemiologists" and the "scientific community" be. I understand the politicians, as they always lie to "keep calm". But scientists should know better! The D/(D+R) might not be "commonly used" but I think in a pandemic with this parameters, it is useful. And if "the Scientific Community" has not used it yet, then they should start doing so. Calculating CFR as Deaths/Confirmed is very misleading to the public. I actually think it's CRIMINAL to mislead the public in this way. The Deaths/Confirmed ratio will keep being artificially depressed as long as the daily growth is high and only show the true figure when the growth tapers off substantially. I think that's VERY BAD! The D/(D+R)will keep coming down and eventually will approach the real CFR, even if the confirmed growth rate keeps being high. Please let D/(D+R) stay. Be patient... it will get closer to the real CFR every day.--AAAAA (talk) 00:03, 7 February 2020 (UTC)[reply]

Misleading entry in case statistics table, D/(D+R), and D/C

The D/(D+R) entry in the case statistics table is very misleading in the sense that it is often interpreted as an estimate for mortality. This is only true if the time delay between infection and death (deltaT_D) is equal to the delay between infection and "recovery" (deltaT_R) e.g. for a given illness, if deltaT_R is considerably longer than deltaT_D then D/(D+R) will be much larger than the actual mortality % and may never even tend fast enough to the real mortality value. As a side note, it would seem appropriate to give the actual definition of "recovered case" as used by the (Chinese) medical officials: does recovered mean no more symptoms ? Undetectable levels of the virus ? ... TheRightKindOfDoctor (talk) 18:27, 2 February 2020 (UTC)[reply]

@TheRightKindOfDoctor: I removed the addition of meaningless columns. As for the "Recovered (cumulative)", it comes from the NHC source "累计治愈出院" which means "Cumulative cured/recovered discharge".―― Phoenix7777 (talk) 21:38, 2 February 2020 (UTC)[reply]

@TheRightKindOfDoctor: You are almost right in your assessment. D/(D+R) should not be interpreted as a real estimate for mortality. However, D/(D+R) will trend towards the real mortality value over time, especially when the infection rate growth slows down.--AAAAA (talk) 00:37, 3 February 2020 (UTC)[reply]

@Phoenix7777: I don't think they are meaningless parameters. They help assess the severity (or not) of the outbreak--AAAAA (talk) 00:37, 3 February 2020 UTC)

@AAAAA: I am sorry to have to insist but D/(D+R) is being misused by others (elsewhere and on FB). It seems a shame that Wikipedia would unwantingly serve as a means to propagate disinformation. D/(D+R) does not show in any way the severity of an outbreak e.g. a below 1% mortality rate with a long recovery time will still show a very high D/(D+R) for months. Even with 9x more recoveries than deaths (this will eventually be the case) D/(D+R) will still be at 10%, regardless of actual mortality rate. Only at the very end of the outbreak will D/(D+R) equate to the mortality ... therefore D/(D+R) truly is meaningless during the outbreak ! TheRightKindOfDoctor (talk) 07:08, 3 February 2020 (UTC)[reply]

@TheRightKindOfDoctor: Let's say the mortality rate is 1% and the recovery time is 2 weeks. Once there are no new infections, after 2 weeks the recoveries should be 99X the deaths. At that time, the D/(D+R) should be equal to the real mortality rate. I understand your point, but on the other side of the argument, many people see 361 Deaths and 17,205 infected and calculated mortality rate of 2.1% and think this outbreak is not scary. Considering 15% of the initial 41 patients have died, seems that mortality rate on this one is about that 15%, unless they find a new treatment. I understand 41 is a low sample size. In any case, I still think D/(D+R) is a good trend indicator. --AAAAA (talk) 15:40, 3 February 2020 (UTC)[reply]

@AAAAA: @TheRightKindOfDoctor: Regarding the initial 41 cases (Jan 2nd) reviewed January 24th paper in the The Lancet. Three weeks later (Jan 22nd) 7 were still hospitalized, 28 had been discharged, and 6 had died. D/(D+R) = 6/(6+28)=18%. I'm not optimistic that 82% (100%-18%) of the 7 unrecovered will recover, so I'd take 18% to be on the low side for this cohort. In the intitial phase of an epidemic, D/(D+R) is a known overestimator, and D/C is a known underestimator. See for example, https://medicalsciences.stackexchange.com/questions/21001/what-is-a-sophisticated-estimate-of-the-2019-ncov-fatality-rate . As is suggested in that link, I suggest a timecourse graph be added to the page. The graph would show the timecourse curves for the estimators D/C and D/(D+R), as well as annotated individual data points, such as (x,y) = (Jan 22, 18%). SailBelow (talk) 18:51, 7 February 2020 (UTC)[reply]

@SailBelow: I agree that we should show the graphs. I don't know how to add graphs that automatically update when the numbers are updated. I would add graphs for: Daily Increase in Confirmed, D/(D+R), D/C (Deaths/Confirmed), (D+R)/C or Resolved/Confirmed. Also... the discussion you quote mentions this "Since both of the naive estimators will yield the same result when the epidemic is over, I suppose one could track both, and extrapolate the curves to where they will meet." Yes, both D/(D+R) and D/C will be exactly the same when the outbreak is over and every single case has been resolved. However, if the daily increase remains high, the D/C will keep being depressed for a long time so it won't really help too much. If days to resolve is for example 15 and daily increase is on average 15%, the number of confirmed cases is 1.15^15-1, which is over 8 times the number of resolved cases. That will keep D/C very low for a long time and that's what we are seeing now. While the D/(D+R) number has been dropping every day, D/C has been steady at 2%. What is sad is that the WHO and the media keep insisting that the risk of death is "Low" at 2% and that's extremely misleading.--AAAAA (talk) 14:13, 8 February 2020 (UTC)[reply]

IMO this number should be removed because it is too prone to misinterpretation. If anyone really desires the number for personal, one can calculate it by themselves using the other data present in table. Džuris (talk) 15:52, 3 February 2020 (UTC)[reply]

@Džuris: There are many tables in Wikipedia that have ratios calculated from the data. I think it's not a bad idea to provide more information. Interpretation should be left to the users. I still think that's a good indicator. It's obviously not the real mortality rate. As TheRightKindOfDoctor mentioned above, there's probably a good delta between the average time from diagnosis to death and the average time from diagnosis to recovery. Once that data is available, another column could be created using the deaths data with n days prior to the recovered data (n being the difference in days between death and recovery). I calculated the numbers with a 7d delta (just a guesstimate) and the rate is around 18-19% for the last 4 days. What is currently misinterpreted by most people and even the media is that they use the Deaths/Infected ratio, which is currently 2.1%, and disregard this serious disease as "low mortality". Based on the current numbers, if a new treatment is not found, seems that the real mortality rate is definitely much higher than 2%... possibly as high as 15%. Also, as days pass, the D/(D+R) will keep declining and should trend towards the real mortality rate.

There is nothing useful that can be interpreted from that data right at this moment. We can include it at a later time, but just not now. Orangewarning (talk) 12:27, 5 February 2020 (UTC)[reply]

@AAAAA: You got something to say, say it here for everyone to discuss. Don't put it in my talk page. Anyway, I'll just address your concerns here. You said that the 2% figure is far from the truth. What is the truth? Because D/(D+R) is equally far from the "truth" too. Now as for whether or not people think that the virus is a big deal, the inclusion of D/(D+R) or CFR for that matter, is not going to change anyone's opinions. The average joe wants someone to tell them what the mortality rate is, they don't know what CFR or D/(D+R) means. If the person knows what it means, then they are not going to be "misled" by the 2% figure in the first place. While you believe that the final CFR will be 8-15%, I would like to remind you that wikipedia articles are meant to be impartial and your edits should not be motivated by what you believe facts to be. While I understand your worries over people not taking this virus seriously, ultimately wikipedia is not a health bulletin, it is not the tool for letting people know how they should or should not act. There are more efficient ways to do it. Honestly, most people won't even be looking at this table. I'm not censoring anything, I'm just removing useless information. And no, I do not work for the chinese government, or any government for that matter. Anyway. I'm just gonna leave it up since you so insist but I still want to reiterate that there is no point in cluttering the table right now. Orangewarning (talk) 11:11, 6 February 2020 (UTC)[reply]

@Orangewarning: Sorry for writing to you in your talk page. I don't know why did I do that in the first place. Yes, D/(D+R) is far from the truth right now, but it is not an invented number... it's just a ratio calculated using numbers currently in the table. About changing people's opinions: I disagree with you. Before I started looking at the numbers of the table and making my own conclusions, I had spoken (by chance) with the Doctor in charge of infections diseases for a large state hospital and she told me that "the fatality rate seems to be low and stable at 2% so not too much worry". And after analyzing the numbers and figuring out that the final CFR will be substantially more than that, it got me worried. And even more that such a Doctor, the WHO and all the media keep thinking that the real CFR is 2% when it's definitely not. Now... about me believing that the final CFR will be 8-15% or whatever, I don't see what's the problem... I didn't post that in the article itself... only here in the talk page. Finally, I appreciate your offer to "leave it up". I think that in a few more days the D/(D+R) will start to stabilize and show something close to the final CFR. Thank you.--AAAAA (talk) 15:02, 7 February 2020 (UTC)[reply]

Dear :@AAAAA:, maybe you can have a look at a web resource on epidemiology. You will see that deaths/cases (CFR) is commonly used (https://wiki.riteme.site/wiki/Case_fatality_rate). Why is it not in the stats table ? Why is D/(D+R) there when nobody in the field uses it ? The average death rate of nCoV is not 15% because it is highly dependent on the age distribution of those initial 41 patients (see for example SARS age-specific CFR (Hong Kong): 0-24 0% (n=0), 25-44 6% (n=29), 45-64 15% (n=35), 65+ 52% (n=87) ) and it's not even close to ANY of the last 12 values of D/(D+R). Alas, in just a few more days, there will have been 500 deaths and ~ 25000 cases (CFR IS quite stable at 2%) but despite about 1250 predicted recoveries, D/(D+R) will still give a rather nonsensical/confusing 28%: from 57.8% to 28% in 1 week ... what does it mean to the average reader ? Bottom line: such (mis)interpretations should NOT be left to the users, not to mention click-bait articles ! TheRightKindOfDoctor (talk) 22:40, 3 February 2020 (UTC)[reply]

@TheRightKindOfDoctor: If you read the article you mentioned (https://wiki.riteme.site/wiki/Case_fatality_rate) closely, the example calculation states: "Assume 9 deaths among 100 people in a community all diagnosed with the same disease. This means that among the 100 people formally diagnosed with the disease, 9 died and 91 recovered. The CFR, therefore, would be 9%." This example is assuming that the disease run its course and there was a final outcome. So in the sample, 100 (D+R), 9 (D) and 91 (R), CFR = D/(D+R). So... the article actually SUPPORTS the current calculation of D/(D+R). The article then states "If some of the cases have not yet resolved (either died or recovered) at the time of analysis, this could lead to bias in estimating the CFR." And I agree... there is bias right now, but that bias will be reduced each day. Stating that the CFR is "quite stable at 2%" is incorrect and very misleading. This table is an example of how the media misunderstands it. It's comparing the CURRENT ratio of Deaths/Confirmed Cases (about 2%) with the FINAL ratio of Deaths/Total Cases of Sars. If you look at this chart, you will see that the CFR of SARS was close to 10%, where initially it was wrongly calculated as 3.5%. I just added a note to the CFR Column Title: "CFR = Case Fatality Rate. This number should only be considered final once every person diagnosed with the disease has either died or recovered". If you have a suggestion to improve the note, let me know. --AAAAA (talk) 13:26, 4 February 2020 (UTC)[reply]

Agree with removing D/(D+R). It does not seem to be common epidemiological description, and seems particularly likely to be misleading at this point. If you want a hint as to why, just note that it's value has changed from 55% to 33% in a week. It takes people longer to recover than to die, so R is a particularly lagging indicator. CFR is also potentially problematic when reported early in an outbreak, but at least that one is fairly standard (and widely reported) and has also been more stable (only going from 2.2% to 2.0% during the last week). Dragons flight (talk) 11:47, 6 February 2020 (UTC)[reply]

@Dragons flight: I think exactly the opposite is true. The "standard" is reporting 2% that that's FAR AWAY from reality. I estimate that right now the CFR is around 10%. The media is parroting the 2% and that gives comfort to most people, when it's a complete LIE when you are talking about confirmed cases growth of 15%-20% daily. and seems it it's taking about 10-15 days on average from diagnosis to resolution (death or recovery). 1.15^15 is over 8. This means that by the time people have died or recovered, when you calculate what people are wrongly calling CFR it gives you the death rate of people infected 10-15 days before and divides is by a mass of people that is 8 times as large that has 7 times the unresolved cases. 2% gives people a false sense of comfort. And even if the real CFR is now 10%, the deaths/confirmed will keep giving that low 2% if the confirmed cases keep growing at 15% or more per day. That's really bad. On the other hand, the D/(D+R) of the resolved cases will keep going down with time, even if the confirmed growth is 15%, and will trend towards the real CFR. It's good to see that it goes down every day, and also is a reminder of how serious this pandemic is.--AAAAA (talk) 23:37, 6 February 2020 (UTC)[reply]

Your concerns are reasonable, but I think you are missing some details. In overwhelmed Hubei province, where most of the deaths have occurred, I think you'll find the delays between the onset of symptoms and the laboratory confirmation of the diagnosis are actually eating up a majority of the time between the onset of symptoms and death for many of the cases. So while you are right to be concerned that deaths lag confirmed cases, I would suggest that you are overestimating the impact of that lag.

Using a 9 day delta(days between average diagnosis to recovery minus days between average diagnosis to deach), we get these D/(D+R) ratios in the last few days: 11.1% 10.5% 11.2% 10.6% 10.3% 9.9%. Using a 10 day delta we get these: 7.1% 7.9% 8.1% 8.2% 8.4% 7.9%. Using a 11 day delta we get these: 4.9% 5.0% 6.1% 5.9% 6.5% 6.4%. And using a 12 day delta we get these: 2.7% 3.5% 3.8% 4.4% 4.6% 4.9%. Based on these numbers I now estimate that the final CFR is probably around 6%.--AAAAA (talk) 14:32, 7 February 2020 (UTC)[reply]

However, the question of whether CFR is a good metric is rather separate to the question of whether D/D+R is a good one. Which I don't think it is given the dynamics currently at play and the very high D/D+R numbers. (BTW, you say that D/D+R will keep going down if the case growth stays at 15%, but that's not true. While cases are growing exponentially, then D and R also grow exponentially, albeit with a lag, and D/D+R will be a constant.) Dragons flight (talk) 00:40, 7 February 2020 (UTC)[reply]

I agree with your first sentence, but I think that naming Deaths/Confirmed as Case Fatality Rate with most cases unresolved is extremely misleading. CFR only makes sense once every case has been resolved. And regarding your statement about the D/(D+R) will be a constant as time passes, you are wrong. First, you can empirically see that the number has gone down every day since Jan 27th. Second, I think you can agree that the Final CFR and the Final D/(D+R) will be equal once every case has been resolved. At such time 100% of Confirmed Cases will have either Died or Recovered, so Final CFR = Final D/ConfirmedCases = Final D/(Died+Recovered). So D/(D+R) will keep going down until it reaches the final CFR, in an asymptotic fashion. But the "daily CFR" that the WHO and all the media keep repeating (2%), could remain at that constant number for a long while, giving a false sense of security. And that's VERY WRONG.--AAAAA (talk) 14:32, 7 February 2020 (UTC)[reply]

During a period of true exponential growth one can show mathematically that both CFR and D/(D+R) would be expected remain constant (assuming a constant mortality rate, and constant lags, etc.). They converge to the same value, but the question is which (if either) should be reported now, when we are far from the end of the outbreak. I don't see much argument for preferring the consistently high estimator of D/(D+R). Dragons flight (talk) 18:08, 7 February 2020 (UTC)[reply]

@AAAAA: You do realize that with a statement such as "I cannot begin to understand how bad can the 'epidemiologists' and the 'scientific community' be." you have lost ALL credibility ? CFR is strictly defined ... there is no true of false CFR. Whether people (readers, politicians, journalists, you,.. etc) understand or not the meaning of CFR is NOT a valid reason for you (or anyone) to try to modify its definition or for you to invent a "new/better" indicator than what the scientific community uses. D/(D+R) is to be removed ASAP. Adding CFR is, on the other hand, completely open to debate. TheRightKindOfDoctor (talk) 16:33, 7 February 2020 (UTC)[reply]

Why is it so bad to criticize the "scientific community"? So scientists are never wrong? Wikipedia itself was strongly criticized by scientists and scholars, especially at the beginning. Look at it now. What I can tell you with 100% accuracy is that the final CFR of an outbreak can only be defined once every single confirmed case has been resolved (either died or recovered). Calculating a CFR in the meantime is extremely misleading. In an outbreak with a high daily increase and long resolution term, the quantity of unresolved cases will severely depress the daily calculated CFR, so it's USELESS... absolutely USELESS. And not only useless but it's misleading as many people in the world might make life or death decisions based on how "comfortable" or worried they are with such CFR. A 2% CFR makes too many people too comfortable and that's WRONG!--AAAAA (talk) 14:32, 7 February 2020 (UTC)[reply]

Should we report that the official case numbers fit R=0.9999 with a quadratic curve?

Since January 27th 2020: https://twitter.com/evdefender/status/1225408294585393153

Case Fatality Rate (CFR) seemed to have been fitted to 2.2%, then 2.1%, then 2.0% for a few days.

E.g. It's probably a model, and not actual case statistics. :(

Henk Poley (talk) 15:01, 6 February 2020 (UTC)[reply]

That thread is crap. You fit 5 or 8 points slowly increasing points with 3 free parameters? Of course, you can get an excellent fit. That's not an appropriate use of statistics, his conclusions are misleading. It certainly isn't appropriate for Wikipedia. Dragons flight (talk) 15:14, 6 February 2020 (UTC)[reply]

Thanks for your opinion. 27 Jan. to 6 Feb. is more like 10 points, btw - Henk Poley (talk) 15:29, 6 February 2020 (UTC)[reply]

I've look at Adjusted R-squared for p=3 and n=10. Which calculates to: 1-(1-0.9999) * ((10-1)/(10-3-1)) = 0.99985 - Henk Poley (talk) 15:54, 6 February 2020 (UTC)[reply]

Henk, here is an example. I generated 10 random numbers from a normal distribution with mean = 500 and standard deviation = 100:

Day	1	2	3	4	5	6	7	8	9	10
Random Value	540	275	455	408	393	455	377	495	580	449
Cumulative	540	814	1269	1678	2071	2526	2904	3399	3978	4428
y = 98.02D^2 + 328.07D + 178.81	512	876	1255	1652	2066	2499	2952	3427	3924	4445

After fitting the quadratic function, I get a correlation, R = 0.9996.

Obviously, a highly significant fit, right? Well, no. Those are just some random numbers with the added effect of transforming the numbers into a cumulative form. Dragons flight (talk) 20:33, 6 February 2020 (UTC)[reply]

@Dragons flight: It all points at the numbers being manipulated by the Chinese Government. But... it is INCORRECT to call it CFR. CFR can ONLY be calculated when the disease has finally run its course. The media and epidemiologists are doing a disservice by repeating that CFR is 2%. Real CFR right now seems o be close to 10% --AAAAA (talk) 23:44, 6 February 2020 (UTC)[reply]

Thanks for the reply. One other checkmark on that you shouldn't do statistics if you don't understand what you are doing. (But where do you get the experience?) Henk Poley (talk) 05:23, 7 February 2020 (UTC)[reply]

@AAAAA: Do you realize that with a statement such as "I cannot begin to understand how bad can the 'epidemiologists' and the 'scientific community' be." you have lost ALL credibility ?!! CFR is strictly defined ... there is no true of false CFR. Whether people (readers, politicians, journalists, you,.. etc) understand or not the meaning of CFR is NOT a valid reason for you (or anyone) to try to modify its definition or for you to invent a "new/better" indicator than what the scientific community uses. D/(D+R) is to be removed ASAP. Adding CFR is, on the other hand, completely open to debate. TheRightKindOfDoctor (talk) 16:33, 7 February 2020 (UTC)[reply]

@TheRightKindOfDoctor: CFR cannot be true or correct unless the outbreak has finished and every confirmed cases resolved (either died or recovered).--AAAAA (talk) 20:34, 8 February 2020 (UTC)[reply]

@AAAAA:. I really dont know if you actually don't understand all this or if it's something else ? Let me try again: CFR is NOT what you seem to think it is. CFR = D(dt)/C(dt) is well defined at ALL time points of an outbreak. NO it is not "true" of "false" depending on what YOU think or believe it should be or on what values YOU think it should take. CFR is NOT the "probability of death" because it is the CASE FATALITY RATIO (or rate, or risk, whatever one prefers). Trivially, only at the the end, when all are either back home or 6 feet under, does CFR match the average probability of death ... that does NOT mean that CFR is FALSE or WRONG or should not be used during an outbreak. (to be continued in the "== 2019-02-07: Proposal for mortality graph with under-estimator curve, over-estimator curve, and early data ==" talk section). TheRightKindOfDoctor (talk) 21:28, 8 February 2020 (UTC)[reply]

@TheRightKindOfDoctor: OK. You win. CFR is whatever you say and whatever people believe it is. But what I find really bad is that the WHO and the media promote the sense that the "Risk of Death" is low because the CFR is currently low. And that's what really bothers me. I think that the worst thing you can do on any situation is LIE to the people. The people have the right to know what's going on, so they can make the best decisions for them and their families. Showing only the CFR, which I think you can agree that it could be a under-estimator, will give a false sense of security to the people of the world. That's not good.--AAAAA (talk) 23:46, 8 February 2020 (UTC)[reply]

@TheRightKindOfDoctor when AAAAA says the CFR is false he is indicated his belief that the numbers are falsified.

@AAAAA If the numbers are false, then we don't know either the numerator or denominator. The CFR can be higher or lower. What we do know, I think you will agree, is that the CFR outside of china is very low. So one could be of the opinion that CFR is over estimated and the risk is lower than you are implying. Daniel.Cardenas (talk) 23:56, 8 February 2020 (UTC)[reply]

@Daniel.Cardenas Alas no ! AAAAA thinks doctors, scientists, WHO, etc ... are all lying to the public by using the CFR indicator. The discussion was not on the reliability/quality of the data provided by China (yes, I'm excluding the possibility of fake data), but on CFR. CFR is NOT whatever I say @AAAAA, it's simply defined in every textbook on epidemiology you can find. TheRightKindOfDoctor (talk) 00:14, 9 February 2020 (UTC)[reply]