Hacker News Reader

Comment by jmnicolas | Hacker News Reader

I can say that for march to may in France everybody that died with COVID like symptoms was counted as a COVID fatality since we didn't have enough tests.

So it could very well be that many people died of the flu but were counted as COVID, we'll never know.

View on HN

But this article is using flu-tests to test for specimens of the flu.

Why would a lack of COVID tests mess up the flu-tests that they're using?

He’s saying the published stats could be off because of a bias toward diagnosing flu deaths as COVID deaths without testing for COVID. That is, the suggestion seems to me to be that the sampling methodology may underrepresent flu. I don’t know that I agree, but I suppose it’s possible.

The idea that 172 different strains of the Flu virus can be accurately detected and tracked, but suddenly adding a 173rd virus (COVID19) to the mix would mess up their survey seems a bit ridiculous to me.

-----

These flu-surveys are extremely important for designing next year's flu vaccine. There are hundreds, if not thousands, of flu strains, since that virus is constantly evolving. But the flu vaccine can only protect against 5ish strains.

Some strains (in particular: Flu TypeA) are known to spread much faster than others (ie: Flu TypeC). So even if TypeC is more common, maybe Flu TypeA is prioritized because of the risk of it becoming a pandemic a year out.

They didn't perform flu tests on all "covid" deaths.

But the vast majority of samples are coming from people who will ultimately survive, but who have respiratory illness. Anyhow, scientists have considered testing bias.

https://www.cdc.gov/mmwr/volumes/69/wr/pdfs/mm6937a6-H.pdf

“Initially, declines in influenza virus activity were attributed to decreased testing, because persons with respiratory symptoms were often preferentially referred for SARS-CoV-2 assessment and testing. However, renewed efforts by public health officials and clinicians to test samples for influenza resulted in adequate numbers tested and detection of little to no influenza virus.”

Most people don't die from flu, so that means it might be underrepresented by 0.1%.

If anything, there's less testing for flu because of covid. But nonetheless, this is sampling, not testing, the WHO performs consistent randomized sampling over time to keep the data meaningful. If they just were based on testing of reported flu cases the data would be heavily biased and basically useless in tracking the disease.

While testing for flu in the global sample is indeed down 21% compared to last year, positive cases are (currently) down 97%.

From their fortnightly report of 28 October 2019:

"data to FluNet for the time period from 30 September 2019 to 13 October 2019. The WHO GISRS laboratories tested more than 102881 specimens during that time period. 5005 were positive for influenza viruses"

And of 26 October 2020:

"...from 28 September 2020 to 11 October 2020. The WHO GISRS laboratories tested more than 81257 specimens during that time period. A total of 172 specimens were positive for influenza viruses"

This is completely irrelevant to the linked data.

Excess mortality data is here: https://www.euromomo.eu/graphs-and-maps/

Looks like minimal deaths from flu during normal flu months in 2020 followed by a large spike (COVID). The spike appears at the end of the flu season, but the previous patterns suggest you should have been seeing excess mortality earlier if flu was bad this year, so my money is on minimal flu deaths.

Reply on news.ycombinator.com