10

This has been asked on medical science SE, but it's probably more suitable here. Worldometers seemingly reports zero deaths from Covid-19 in China in 2021:

enter image description here

But the WHO's similar graph tells a different story for the outbreaks this summer:

enter image description here

So why is there this discrepancy? Do the WHO's death estimates differ from official Chinese data?

Based on a quick "visual scan" of China CDC data it seems indeed they have reported zero Covid-19 deaths in 2021. So, I guess the question is: why does the WHO disagree with that and how did the WHO estimate the number of deaths in 2021 in China?

Fizz
  • 57,051
  • 18
  • 175
  • 291
  • Looks like data error. European Centre for Disease Prevention and Control disagrees: https://www.ecdc.europa.eu/en/geographical-distribution-2019-ncov-cases – pinegulf Oct 15 '21 at 07:34
  • @pinegulf: not sure what you mean. – Fizz Oct 15 '21 at 08:02
  • The source linked reports 452 infected between weeks 39 and 40 2021. Thus I assume the reporting site data flow might be broken. – pinegulf Oct 15 '21 at 08:08
  • I've not time to look into it now, but I remember hearing the phrase recently: "if even ten percent of the mobile phone contracts de-registered in that period were as a result of covid deaths then...", the implication being extrapolation and educated guesswork in the absence of reported data. – Jiminy Cricket. Oct 15 '21 at 08:23
  • 1
    @pinegulf: again, I'm not sure why that is a problem/contradiction. I'm asking about deaths, not number of infected. Please write an answer instead of these cryptic comments. (The number of *infected* in 2021 is not zero on worldometers; their infected graph is horrible [too], but if you mouse over it, it has non-zero daily infections in 2021 for a good number of days. The ECDC instead reports those infections weekly instead of daily, so those numbers are expect to be different. The ECDC doesn't seem to report deaths from China except cumulatively since the begging of the pandemic.) – Fizz Oct 15 '21 at 08:31
  • 1
    This site (https://www.statista.com/statistics/1092918/china-wuhan-coronavirus-2019ncov-confirmed-and-deceased-number/) reports 4,636 deaths in China up to Apr. 17, 2020 and then the number of deaths increases slowly reaching 4782 on January 1, 2021, and 4849 on October 10, 2021. So, in conclusion, only 4849 - 4782 = 67 people died in China of Covid-19 in 2021 (up to Oct. 10, 2021) ! – Robert Werner Oct 15 '21 at 16:48
  • The Worldmeters site certainly has a problem. Note that the Worldmeters site also has a graph showing total deaths, you can hover that and see numbers. While the numbers do not agree with the WHO numbers they do show 3 deaths since China went to stomping out hotspots. However, their daily graph doesn't show the associated blips. – Loren Pechtel Oct 16 '21 at 23:18
  • @Fizz Note that with weekly reporting China will show less than one death per day and a simple roundoff problem could cause this. Living in a state with some very sparsely populated counties I've seen quite a few artifacts that are probably roundoff related. – Loren Pechtel Oct 16 '21 at 23:20
  • 1
    @Loren Pechtel , I do not understand what is the difference between 0, 1, 2 and 67 deaths in China for the entire 2021 up to the present (see my previous comment). China, with a population of 1402 mil people, has at most 67 deaths/41.3 weeks = 1.62 deaths / week or 0.00115 deaths/week/mil people !!! which is zero! Yes, China reported 0 deaths/mil/week and nobody else reported more about China with a very good approximation. The one who opened this topic asked a quite pertinent question. and the answer by all accounts is affirmative and at the same time incredible. – Robert Werner Oct 17 '21 at 01:45
  • @RobertWerner I'm suggesting that this might just be a data artifact from careless rounding. – Loren Pechtel Oct 17 '21 at 01:54
  • @LorenPechtel: CCDC has raw numbers, not per capita, so I don't see what rounding is involved. – Fizz Oct 17 '21 at 01:57
  • @Fizz I'm not talking per capita, I'm talking per day. – Loren Pechtel Oct 17 '21 at 02:33
  • @LorenPechtel: the CCDC doesn't have anything like moving averages, so again I don't see rounding being an issue. – Fizz Oct 17 '21 at 02:37
  • 1
    If you click on “January 26” on the China CDC data link in the question, it says “1 new death (domestic case, Jilin province)” from the previous day (Jan 25). – Rebecca J. Stones Oct 26 '21 at 09:37
  • @RebeccaJ.Stones: Interesting. Indeed it says "1 new death (domestic case, Jilin Province)" in the details... but the summary section has "Deaths: 0 new." for whole of January (every day from Jan 2 to Jan 31). Actually, the same is the case for Jan 14; in the details it has "1 new death (domestic case, Hebei Province)". – Fizz Oct 26 '21 at 10:01

2 Answers2

4

The difference seems to come from using different sources. From the wordometer website:

For the COVID-19 data, we collect data from official reports, directly from Government's communication channels or indirectly, through local media sources when deemed reliable.

The WHO on the other hand collects their own data. I haven't found a detailed explanation but they quote their data with

Source:World Health Organization

This will involve official government communication but that is surely not their only source.

Note that different sources may use different counting methods or have different methology otherwise so different numbers may mean that one source is wrong but that doesn't have to be the case.

quarague
  • 1,474
  • 9
  • 13
  • 1
    Yes, obviously WHO isn't just relying on the stream of zeros from CCDC, but that observation is rather restating the question. – Fizz Oct 15 '21 at 11:03
  • 2
    @Fizz The point is more that wordometer is exclusively relying on the data from the Chinese government. The WHO is not. – quarague Oct 15 '21 at 11:50
  • The European Centre for Disease Prevention and Control and Worldometers say basically the same thing. The discrepancy is small, negligible. The situation of Covid-19 deaths in China up to October 15, 2021: (Source 1) Total deaths - 4,636 (see: https://www.worldometers.info/coronavirus/country/china/); (Source 2) Total deaths - 4849 (see: https://www.ecdc.europa.eu/en/geographical-distribution-2019-ncov-cases). 4849-4636=213. The difference between the reports on ecdc and worldometers is insignificant. – Robert Werner Oct 15 '21 at 16:29
  • 1
    @Fizz: The question asks for the reason of the discrepancy between the two figures. This answer points out that the two figures probably use two different data sources. What else is there to answer? If you feel this explanation lacking, there might be a problem with the question instead. – Schmuddi Oct 16 '21 at 07:22
  • @Schmuddi: I don't see only a "minor discrepancy" between a stream of absolute zeroes (CCDC) and some non-zero death values, raising to the hundreds at one point (WHO). It's only a "minor discrepancy" if you take the sum over the whole pandemic in China. But for 2021, it's not minor discrepancy, IMHO. – Fizz Oct 16 '21 at 07:24
  • @Fizz, I agree, and I've immediately edited that part of my comment after posting. – Schmuddi Oct 16 '21 at 07:25
4

I found two Covid-19 deaths in mainland China in 2021 which were both reported by the Chinese government and mentioned in Chinese news.

The discrepancy in the two plots likely arises from the sources having different definitions of "China". Taiwan had an outbreak in June 2021, but these data are counted separately on Worldometer whereas they are counted for China by the WHO (hover over Taiwan on this map). (Note that 187 deaths in the WHO screenshot spans a week.)

Hebei province, reported 14 January 2021:

1月13日0—24时，31个省（自治区、直辖市）和新疆生产建设兵团报告新增确诊病例138例，其中境外输入病例14例（上海8例，广东3例，北京1例，河南1例，广西1例），本土病例124例（河北81例，黑龙江43例）；新增死亡病例1例，为本土病例（在河北）；无新增疑似病例。
Gov.cn, 14 January 2021.

Here's a corresponding video news report:

screenshot of news video
河北新增1例死亡病例详情公布，系中国内地近8个月以来首次增加, 14 January 2021.

Jilin province, reported 26 January 2021:

1月25日0—24时，31个省（自治区、直辖市）和新疆生产建设兵团报告新增确诊病例82例，其中境外输入病例13例（上海8例，广东2例，福建1例，湖南1例，陕西1例），本土病例69例（黑龙江53例，吉林7例，河北5例，北京2例，上海2例）；新增死亡病例1例，为本土病例，在吉林；新增疑似病例3例，均为境外输入病例（均在上海）。
Gov.cn, 26 January 2021.

Here's an example news article about this case.

I didn't find any other cases reported. Searching for site:www.gov.cn "死亡病例1例" after:2021-01-01 only gives these two cases.

Note the linked English CDC China Weekly doesn't list these, but if you click through, they're both listed:

National Health Commission Update on January 14, 2021

..., 1 new death (domestic case, Hebei Province), ...

National Health Commission Update on January 26, 2021

..., 1 new death (domestic case, Jilin Province), ...

Rebecca J. Stones
  • 7,948
  • 6
  • 37
  • 68