China COVID-19 case data now includes asymptomatic cases

[CSSEGISandData]

Hello all,

We are excited to announce that we have finally been able to integrate asymptomatic COVID-19 cases into our data for China and back corrected these values to March 30, 2020 in #5573. The China National Health Commission does not publish cumulative asymptomatic cases, so we have used the following logic to calculate current and historical asymptomatic cases:

Total asymptomatic = "Total from previous day"+"daily new asymptomatic cases"-"cases converted from asymptomatic to confirmed"

The first NHS report published on March 31, 2020 stated there were 128 asymptomatic cases. We have used this as our first entry. Any question please let us know.

Clarification: All cases have been added to the "Unknown, China" entry in the dataset.

4/4/22:

We have found cumulative asymptomatic cases for March 30, 2020 (1541). These cases have been integrated into the time series and back corrected in #5581.

[Artoria2e5]

I think there is a point in clarifying that "asymptomatic" isn't quite... asymptomatic. This is probably of less concern to the data-consumers at large here, but does mean that hearing about one in your building during lockdown is a tiny bit scarier. (Or it can mean that "transmission by asymptomatic people" is not as unexplainable as it sounds. It's all about perspective.)

Let's quote the covid guidelines, ninth edition, translation by deepL:

(B) Diagnostic criteria.

1. Suspected cases.
   Any one of the following epidemiological history and any two of the clinical manifestations.
   If there is no clear epidemiological history, meet 3 of the clinical manifestations; or meet any 2 of the clinical manifestations, and at the same time positive for novel coronavirus-specific IgM antibody (recent vaccination against novel coronavirus is not used as a reference indicator).
   (1) Epidemiological history [very normal stuff, omitted]
   (2) Clinical manifestations
   1 Clinical manifestations related to novel coronavirus pneumonia such as fever and/or respiratory symptoms
   2 Imaging features of novel coronavirus pneumonia as described above;
   3 normal or reduced total white blood cell count and normal or reduced lymphocyte count in the early stage of the disease.
2. Confirmed cases. Suspected cases have one of the following pathogenic or serological evidence: (1) positive nucleic acid test for novel coronavirus; (2) positive novel coronavirus-specific IgM antibody and IgG antibody in those who have not received novel coronavirus vaccine.

What is notable is that:

• Unlike with WHO def, a confirmed case needs to be both RNA-positive and suspected. This is the issue taken care of by this announcement and I must applaud the CSSE for getting that cleaned up.
• There is no real definition of what an asymptomatic case is. If we take it as "positive but not confirmed", that means it would include people who meet only one of the clinical manifestations. That's not exactly asymptomatic in the normal sense of the word.

[pdg137]

Hello,

Could you start using asymptomatic case numbers for the individual provinces, too? These are given in the reports you link to but apparently not incorporated into your files. As a result, the graphs given in Google searches for China often look incorrect and inconsistent. For example, a search for shanghai covid cases returns this graph:

Whereas the overall China graph looks like this:

...making it look like 90% of China's COVID cases are somewhere outside Shanghai, exactly the inverse of the current situation.