4 ms·

Virginia Juked Its Covid-19 Data

23 points by danso 6y ago

angry_octet 6y ago
This has the same feeling as officials in the Soviet Union faking production numbers. What a tragedy of greedy ineptitude.
najarvg 6y ago
Does an institution like CDC have the mandate to ask for the antibody and viral tests to be reported separately by all states? I would seem to me that we would all benefit as a country if this were done and the data were made available freely (raw and reports data)
SpicyLemonZest 6y ago
I dunno, I'm kinda on Virginia's side here. If "pandemic response efforts assumed clean, standard, accessible data would exist", that seems like a fault of the pandemic response efforts, not the people who produce messy and non-standard data. (Important context: Virginia stopped aggregating antibody tests as soon as they learned it was misleading people. https://www.washingtonpost.com/local/virginia-antibody-covid-19-tests-northam-reopening/2020/05/14/fa9f62b0-95e4-11ea-82b4-c8db161ff6e5_story.html https://www.washingtonpost.com/local/virginia-antibody-covid...)
- NathanKP 6y ago
  On the one hand I want to apply Hanlon's Razor ("Never attribute to malice that which is adequately explained by stupidity"). But on the other hand there have been a lot of very dubious choices made through out this entire pandemic, choices where surely there must have been at least one person who objected but was overruled for some reason. Even NYC fudged its numbers early on by choosing to initially only report deaths where the infected person died in a hospital, after getting a positive test result. They were not initially counting people who chose to die at home, died in nursing homes, or died in hospitals before they were able to get one of the limited tests that were available. At some point when you look at the how the numbers are calculated and it is painfully obvious that the numbers are being calculated wrong you have to ask "how did this happen?" Surely there was at least one intelligent person along the way in Virginia who must have suggested "hey let's not mix these test numbers together" or in NYC at least one intelligent person who said "hey let's also count nursing home deaths". When it takes a public expose pointing out how wrong the numbers are in order to get the numbers corrected, that is a problem. It might be just a major case of brain drain in these public service roles, or it might be something more sinister.
  - SpicyLemonZest 6y ago
    But I don't think it's right to call these "fudged". It strikes me as the appropriate division of labor. Government administrators are responsible for reporting the data they've collected; it's pandemic modelers, not administrators, who are best equipped to determine which data is useful and what the data implies. Pandemic modelers realized they needed data on PCR tests only, so they told Virginia that and now they're getting it - the system worked! It'd be a different story if Virginia said "this is the number of test suitable for contact tracing we're doing", or NYC said "this is an unbiased estimate of the total mortality burden of covid-19 in our city". But I don't think they did.
kgin 6y ago
Did anyone else not have any idea what juked means?
- LandR 6y ago
  I've only ever heard it used in a fighting sense, like dodging or feinting a blow. Apparently it also means fake, so they faked their data. But this doesn't seem right either reading the actual article. IT wasn't faked so much as it was aggregated poorly to mislead. Or weren't clear what the data actually showed (antibodies v active infections). Seems a strange choice of word.
m0llusk 6y ago
This seems like alarmism. All of these tests are useful, and the antibody tests scale better as they are quicker and cheaper which makes them useful for contact tracing containment. Because most COVID-19 infections that have negative outcomes, that is hospitalization or death, are in people with one or more other major chronic health problems the numbers are complex and identifying a single cause of death is not necessarily correct. In any case most analysts agree that if you want really good numbers for comparison between countries and over time then the only way to go is with all cause mortality. That number captures all of the most serious impacts and has a strongly identifiable spike.

23 points by danso 6y ago

This has the same feeling as officials in the Soviet Union faking production numbers. What a tragedy of greedy ineptitude.

najarvg 6y ago

Does an institution like CDC have the mandate to ask for the antibody and viral tests to be reported separately by all states? I would seem to me that we would all benefit as a country if this were done and the data were made available freely (raw and reports data)

SpicyLemonZest 6y ago

I dunno, I'm kinda on Virginia's side here. If "pandemic response efforts assumed clean, standard, accessible data would exist", that seems like a fault of the pandemic response efforts, not the people who produce messy and non-standard data. (Important context: Virginia stopped aggregating antibody tests as soon as they learned it was misleading people. https://www.washingtonpost.com/local/virginia-antibody-covid-19-tests-northam-reopening/2020/05/14/fa9f62b0-95e4-11ea-82b4-c8db161ff6e5_story.html https://www.washingtonpost.com/local/virginia-antibody-covid...)

NathanKP 6y ago
On the one hand I want to apply Hanlon's Razor ("Never attribute to malice that which is adequately explained by stupidity"). But on the other hand there have been a lot of very dubious choices made through out this entire pandemic, choices where surely there must have been at least one person who objected but was overruled for some reason. Even NYC fudged its numbers early on by choosing to initially only report deaths where the infected person died in a hospital, after getting a positive test result. They were not initially counting people who chose to die at home, died in nursing homes, or died in hospitals before they were able to get one of the limited tests that were available. At some point when you look at the how the numbers are calculated and it is painfully obvious that the numbers are being calculated wrong you have to ask "how did this happen?" Surely there was at least one intelligent person along the way in Virginia who must have suggested "hey let's not mix these test numbers together" or in NYC at least one intelligent person who said "hey let's also count nursing home deaths". When it takes a public expose pointing out how wrong the numbers are in order to get the numbers corrected, that is a problem. It might be just a major case of brain drain in these public service roles, or it might be something more sinister.
- SpicyLemonZest 6y ago
  But I don't think it's right to call these "fudged". It strikes me as the appropriate division of labor. Government administrators are responsible for reporting the data they've collected; it's pandemic modelers, not administrators, who are best equipped to determine which data is useful and what the data implies. Pandemic modelers realized they needed data on PCR tests only, so they told Virginia that and now they're getting it - the system worked! It'd be a different story if Virginia said "this is the number of test suitable for contact tracing we're doing", or NYC said "this is an unbiased estimate of the total mortality burden of covid-19 in our city". But I don't think they did.

kgin 6y ago

Did anyone else not have any idea what juked means?

LandR 6y ago
I've only ever heard it used in a fighting sense, like dodging or feinting a blow. Apparently it also means fake, so they faked their data. But this doesn't seem right either reading the actual article. IT wasn't faked so much as it was aggregated poorly to mislead. Or weren't clear what the data actually showed (antibodies v active infections). Seems a strange choice of word.

m0llusk 6y ago

This seems like alarmism. All of these tests are useful, and the antibody tests scale better as they are quicker and cheaper which makes them useful for contact tracing containment. Because most COVID-19 infections that have negative outcomes, that is hospitalization or death, are in people with one or more other major chronic health problems the numbers are complex and identifying a single cause of death is not necessarily correct. In any case most analysts agree that if you want really good numbers for comparison between countries and over time then the only way to go is with all cause mortality. That number captures all of the most serious impacts and has a strongly identifiable spike.