Thread by @gdemaneuf

1/ This thread discusses my recent study on the number of 2019 onset cases and its suppression by the Chinese authorities.

That study ended up validating the SCMP leak of gov data (257 Dec 19 cases), with an implication for Nov 19 cases.

PDF: bit.ly/2019Cases

2/ That study was mentioned in a recent Washington Post Opinion from the Editorial Board, available here:
wapo.st/3OuJQg8
@thedeadhandbook

A companion spreadsheet is here:
bit.ly/3S7jAsQ

3/ Before we proceed:

The numbers here are all traceable to data from peer-reviewed papers from very experienced Chinese teams, one of them directly involved in the cleaning and populating of the official database for the CDC.

You could not ask for better sources of data.

4/ Let's start by the main findings about the 2019 cases:

Based on these peer-reviewed papers, there were necessarily 247 to 260 cases (either confirmed or diagnosed) with onset in 2019 in the official Chinese CDC database by end Feb 2020

👉🏻 NOT the 174 of the final WHO report.

5/ also:

The most likely number is toward the high end of 260.
The 247 to 260 include at least 165 laboratory confirmed cases.

👉🏻 NOT the 100 of the final WHO report.

6/ also:

33 of these 2019 onset cases died by the 24 Feb 2020, including 29 lab-confirmed (and 4 diagnosed ones)

Which is much more than what is generally known (15 deaths, from the CDC Weekly report published on 17 Feb 2020).

7/ This nicely matches the South China Morning Post leak of 257 cases for Dec 2019 published 13 March 2020 (which also included an additional 9 November 2019 cases, either confirmed or diagnosed).

👉🏻 The SCMP leak is effectively validated.
h/t @scmp_joma
archive.ph/wip/FeOyq

8/ The SCMP leak is quite detailed (and also follows a near perfect exponential growth).

It is worth quickly giving you the full picture:

9/ So, at this stage, most people will ask 'Who are these conspiracy theorists coming up with silly numbers'?

1st team led by WuChun Cao, well known PLA scientist and epidemiologist who collaborated with Holmes on some pangolin paper, did conferences with Baric, Shi Zhengli, etc

10/ Here is his key paper.
mdpi.com/1660-4601/17/1…

Check in particular Figure 1 in the paper.

This is a figure of laboratory confirmed cases using the official CDC database at 27 Feb 2020, aggregated per week:

11/ The key numbers are not in the text, but they are in the graph.

Let me zoom a bit.

146 = 9 + 41 + 2 + 94 (laboratory) confirmed cases to 29 Dec 2019.

12/ 2nd team led by Chunhua Yu, prof. of Biostatistics at Wuhan Uni and specifically in charge of back-populating and cleaning the official database of SARS-CoV-2 cases with the CDC.

He put out 3 papers with relevant 2019 case numbers.

Main one here:
ghrp.biomedcentral.com/articles/10.11…

13/ Check in particular Table 2, which uses data in the official database as of 24 Feb 2020 and shows:

- 33 deaths for Dec 2019 onset cases (from 29 confirmed & 4 diagnosed cases)

- 135 confirmed cases (29+106)

- 92 diagnosed ones (4+88)

ghrp.biomedcentral.com/articles/10.11…

14/ Now by using Fig. 2 of ghrp.biomedcentral.com/articles/10.11…’ (another Chuanhua's paper which uses the same data), you can even get the onset dates of these 33 deaths.

Note the 2 deaths for 1st Dec 2019 onset.

15/ [Side note:

There was also a weird thing going on with two Dec cases in the official CDC Weekly update on 17 Feb 2020 (using data as of 11 Feb).

The table in that update gave 104 cases, but the graph starting 8 Dec 2019 only had 102. So two cases went missing somewhere.]

16/ With all this, you could just:

- take the 146 confirmed of Wu-Chun Cao et al to 29 Dec 2019,

- fill with the 19 confirmed for 30/31 Dec 2019 from the final WHO report

- add the 92 diagnosed of Chuanhua Yu et al for Dec 2019

Let's see:
146 + 19 + 92 = 257

@Peterfoodsafety

17/ That is exactly the number of Dec 2019 cases reported in the South China Morning Post leak of 13 March 2020.

All by using peer-reviewed papers from top Chinese teams, with excellent access to the CDC, with CDC data extracted at end Feb 2020.

18/ Because of the small mismatch of database extract dates between Wu-Chun Cao and Chuanhua Yu papers (+ final WHO report), I am careful and prefer to show instead that we have a necessary range of 247 to 260 cases at 27 Feb 2020, with the actual number most likely toward 260:

19/ The full derivation of this necessary range of 247 to 260 takes into account the definitions of confirmed and diagnosed and their behaviours over time.

The paper goes through it in detail. No need for me to repeat it.
Check in particular Boxes 1 and 5 if you are interested.

20/ That is enough technicalities already for a first thread about the real number of December 2019 cases from hospital cases, as was recorded in the CDC database at end Feb 2020.

Let me now quickly explain what effectively happened and what it practically means.

21/ Chuanhua’s team and the CDC had only about 3 weeks starting from around 6 Feb 2020 to do their main retrospective search (which also had to focus on retrospective diagnosed cases for 2020), before it got stopped by the gag order imposed on the CDC on 25 Feb 2020.

22/ These must have been very busy times:

They had to review medical files to identify cases according to the just introduced diagnosed definition (most of them over Jan 2020).

They also had to try to get some original samples tested, while also trying to cover 2019 onsets.

23/ And they had to check for duplicate entries, partial records and other data issues.

In the middle of a lockdown.

After a few days of work, the CDC published its key update, on 17 Feb 2020, using their very preliminary data for 2019 as of 11 Feb.
weekly.chinacdc.cn/en/article/doi…

24/ That’s the 104 confirmed cases, which effectively became the anchoring point of the data released to the WHO one year later (100 confirmed cases + 74 diagnosed).

From 16 to 25 Feb, they could focus more on confirmed cases for 2019, as the diagnosed category was then closed.

25/ But everything stopped shortly after the gag order on the CDC on 25 Feb 2020.

By that time, Chuanhua's team and the CDC had nevertheless managed to identify 247 to 260 cases for 2019 onset from hospital records (as at least 165 confirmed, the rest diagnosed).

26/ Most likely the number is indeed 257 as per the SCMP leak a few days later (13 Mar 20), with indeed another 9 Nov 19 cases (either diagnosed or confirmed)

Had they had more time, they would certainly have identified more hospital cases, especially milder ones.

27/ Clearly someone was unhappy to see the CDC work being gagged.

The CDC had just issued that useful update in mid Feb and now they were told to shut up while their database (actually under the control of the Party Secretary of the CDC) was being rolled back to around 15 Feb.

28/ Hence the leak to the SCMP within a few days.

But many people still don’t realise that China did not lift a finger to do any retrospective search after 25 Feb 2020.

Nor that its supposed new retrospective search in Jan 2021 with its 'all-negative' makes no sense.

29/ If China authorities lifted a finger it was to press the delete key on the keyboard, as they effectively rolled the official database back to the (one and only) CDC update of mid February 2020.

That’s the simple and sad story.

30/ Now, with at least 165 lab-confirmed hospital cases instead of the 100 presented to the WHO, @MichaelWorobey should do the honourable thing and retract his Science paper, @ScienceMagazine

It is not his fault if he has been lied to.

Even if he ignored the possibility.

31/ And despite being amply warned.

32/ Also don't imagine that it is only China that decided to play with data to kill some hypothesis.

In February 2020, Western scientists were doing exactly the same thing, clinging desperately to the debunked pangolin straw in Proximal Origins.

33/ Effectively February 2020 started near 3 solid years of lies about the origins.

In China. In the US. In Europe.
Everywhere.

Thanks to #DRASTIC and some FOI emails, we now know better.

34/ I hope you've found this thread helpful.

Follow me @gdemaneuf for more.

Like/Retweet the first tweet below if you can: