Page 1 of 3 123 LastLast
Results 1 to 10 of 26

Thread: A statistic of interest.

  1. #1
    Professor

    Join Date
    May 2019
    Location
    Staten Island, NY USA
    Last Seen
    Today @ 06:49 PM
    Gender
    Lean
    Other
    Posts
    1,969

    A statistic of interest.

    There are a number of quantities which are of importance in tracking the progress of the COVID-19 zoonotic pandemic. These include the number of people who have tested positive for the virus in a given population, the number of deaths in that population, the number of recoveries, and others.

    One which seems to be of great interest is the rate of mortality. To put it in simple words, just how deadly is the disease? We've quite a range of estimates.

    The rate of mortality has another use. It can be applied to give an estimate of how we are doing in our program to test Americans for the disease. Here's how it goes.

    One estimate of the mortality rate, with obvious flaws, is that given as a percent of those who have tested positive for the disease. During the period 3/22-4, it hovered at its lowest point, around 1.2%. As of today, 4/6, it stands at 2.86% with 9,620 recorded virus-caused deaths and 336,851 recorded positives. If we boost the number of people who test positive from its present value to bring the mortality rate back to 1.2%, we would need 801,670 confirmed cases.

    That means that we may be behind in testing by 464,800 tests as compared to the 3/22-4 period.

    [Disclaimer: I'm at home with statistical techniques both through graduate study and professional use. I've been tracking these and other data since 3/18/2020.]
    Last edited by Torus34; 04-06-20 at 06:59 AM.
    "And gladly wolde he lerne and gladly teche." Chaucer, the Canterbury Tales.

  2. #2
    global liberation


    ecofarm's Avatar
    Join Date
    Jun 2010
    Location
    Miami
    Last Seen
    Today @ 10:54 PM
    Gender
    Lean
    Independent
    Posts
    112,446

    Re: A statistic of interest.

    It is theoretically possible to derive the number of cases from the number of deaths; however, one cannot assume all test results are positive. As some tests are negative, merely adding a number of tests and presuming they are all positive does not arrive at the number of tests that should have been conducted.

  3. #3
    Sage
    longview's Avatar
    Join Date
    Jan 2012
    Location
    Texas
    Last Seen
    Today @ 07:42 AM
    Gender
    Lean
    Conservative
    Posts
    24,979

    Re: A statistic of interest.

    Quote Originally Posted by Torus34 View Post
    There are a number of quantities which are of importance in tracking the progress of the COVID-19 zoonotic pandemic. These include the number of people who have tested positive for the virus in a given population, the number of deaths in that population, the number of recoveries, and others.

    One which seems to be of great interest is the rate of mortality. To put it in simple words, just how deadly is the disease? We've quite a range of estimates.

    The rate of mortality has another use. It can be applied to give an estimate of how we are doing in our program to test Americans for the disease. Here's how it goes.

    One estimate of the mortality rate, with obvious flaws, is that given as a percent of those who have tested positive for the disease. During the period 3/22-4, it hovered at its lowest point, around 1.2%. As of today, 4/6, it stands at 2.86% with 9,620 recorded virus-caused deaths and 336,851 recorded positives. If we boost the number of people who test positive from its present value to bring the mortality rate back to 1.2%, we would need 801,670 confirmed cases.

    That means that we may be behind in testing by 464,800 tests as compared to the 3/22-4 period.

    [Disclaimer: I'm at home with statistical techniques both through graduate study and professional use. I've been tracking these and other data since 3/18/2020.]
    We may never really know the mortality, unless we know how many people actually contracted Covid-19, and we may not know that unless we just assume 100% were exposed.
    I think the Mayo clinic is working on an antibody test to look at who is making antibodies, I.E. who has gotten and recovered from Covid-19.
    In my County, I would be surprised if 1% of the population has been tested.

  4. #4
    Professor

    Join Date
    May 2019
    Location
    Staten Island, NY USA
    Last Seen
    Today @ 06:49 PM
    Gender
    Lean
    Other
    Posts
    1,969

    Re: A statistic of interest.

    The above statistic stands in contrast to the plethora of reports and statements of how many test kits are being manufactured and shipped. There appears to be a tight spot in the flow from manufacturer to those actually administering the tests. Whether it be hold-ups in moving merchandise, guidelines which restrict their use or some other constriction is not known to this poor old country mouse. It is doubtful if we will get any answer soon. Vituperative finger-pointing, however, seems to be freely available.
    "And gladly wolde he lerne and gladly teche." Chaucer, the Canterbury Tales.

  5. #5
    Professor

    Join Date
    May 2019
    Location
    Staten Island, NY USA
    Last Seen
    Today @ 06:49 PM
    Gender
    Lean
    Other
    Posts
    1,969

    Re: A statistic of interest.

    Quote Originally Posted by ecofarm View Post
    It is theoretically possible to derive the number of cases from the number of deaths; however, one cannot assume all test results are positive. As some tests are negative, merely adding a number of tests and presuming they are all positive does not arrive at the number of tests that should have been conducted.
    Hi, and thank you for taking time to read and to comment. In the particular statistic detailed in the OP, false positives [Ed.: and false negatives, dude,] can be assumed to 'wash'. they exist in the same proportion throughout, unless there has been a change in test procedure with a significant difference in false results between the tests. There is no information either for or against such a change.

    Regards, and stay well.
    "And gladly wolde he lerne and gladly teche." Chaucer, the Canterbury Tales.

  6. #6
    Professor

    Join Date
    May 2019
    Location
    Staten Island, NY USA
    Last Seen
    Today @ 06:49 PM
    Gender
    Lean
    Other
    Posts
    1,969

    Re: A statistic of interest.

    Quote Originally Posted by longview View Post
    We may never really know the mortality, unless we know how many people actually contracted Covid-19, and we may not know that unless we just assume 100% were exposed.
    I think the Mayo clinic is working on an antibody test to look at who is making antibodies, I.E. who has gotten and recovered from Covid-19.
    In my County, I would be surprised if 1% of the population has been tested.
    Hi! Thanks for the response.

    With time, and with the increase in data which will occur, we'll be able to put tighter and tighter limits on the actual mortality value. It's been asserted by some that as we test more people, the mortality percent figure will drop. That hasn't been borne out by the actual reported data. That, in turn, led to the OP and the tentative conclusion.

    Regards, and stay well.
    "And gladly wolde he lerne and gladly teche." Chaucer, the Canterbury Tales.

  7. #7
    Professor
    Join Date
    Oct 2019
    Last Seen
    06-04-20 @ 02:34 PM
    Gender
    Posts
    1,267

    Re: A statistic of interest.

    South Korea is the best place to get information about coronavirus. They are the only country that has had mass testing and tracking, besides China, whose numbers I don't trust. Here are lots of numbers. Have fun twisting your brain into knots.

    South Korea Coronavirus: 10,284 Cases and 186 Deaths - Worldometer

  8. #8
    Sage
    longview's Avatar
    Join Date
    Jan 2012
    Location
    Texas
    Last Seen
    Today @ 07:42 AM
    Gender
    Lean
    Conservative
    Posts
    24,979

    Re: A statistic of interest.

    Quote Originally Posted by Torus34 View Post
    Hi! Thanks for the response.

    With time, and with the increase in data which will occur, we'll be able to put tighter and tighter limits on the actual mortality value. It's been asserted by some that as we test more people, the mortality percent figure will drop. That hasn't been borne out by the actual reported data. That, in turn, led to the OP and the tentative conclusion.

    Regards, and stay well.
    We are still in the phase where someone must pass a screening process even to be tested,
    but we know there are people out there who get this virus, and the symptoms are mild, and may never ask to be tested,
    or not even know they are sick.
    True mortality would be the number of deaths over the number of infected, we may never know the number of infected.
    I am not sure we even have a good idea of the percentage of people who are asymptomatic, when they get this virus.

  9. #9
    Professor

    Join Date
    May 2019
    Location
    Staten Island, NY USA
    Last Seen
    Today @ 06:49 PM
    Gender
    Lean
    Other
    Posts
    1,969

    Re: A statistic of interest.

    Quote Originally Posted by Digger View Post
    South Korea is the best place to get information about coronavirus. They are the only country that has had mass testing and tracking, besides China, whose numbers I don't trust. Here are lots of numbers. Have fun twisting your brain into knots.

    South Korea Coronavirus: 10,284 Cases and 186 Deaths - Worldometer
    Hi! The South Korea data gives a value of 1.8%. That's not far from the 2.1% I selected as a working 'base line' for the OP. As an aside, had I selected 1.8% or 2.4%, the thrust of the argument would be unchanged. The data suggests we're falling behind in testing.

    Regards, and stay well.
    "And gladly wolde he lerne and gladly teche." Chaucer, the Canterbury Tales.

  10. #10
    global liberation


    ecofarm's Avatar
    Join Date
    Jun 2010
    Location
    Miami
    Last Seen
    Today @ 10:54 PM
    Gender
    Lean
    Independent
    Posts
    112,446

    Re: A statistic of interest.

    Quote Originally Posted by Torus34 View Post
    Hi, and thank you for taking time to read and to comment. In the particular statistic detailed in the OP, false positives [Ed.: and false negatives, dude,] can be assumed to 'wash'. they exist in the same proportion throughout, unless there has been a change in test procedure with a significant difference in false results between the tests. There is no information either for or against such a change.

    Regards, and stay well.
    What I mean is this:

    We have number of deaths. You are using that to establish the number of cases. In theory, that can work (depending on sample size and thus empirical data). So I'll assume it does work (not guaranteed, it's pretty complex and fraught with opportunity for error). We'll say you have arrived at the correct number of cases.

    What you cannot then do is simply add that number to the number of tests for the total tests that should have been conducted. You see, for every positive test result, there are negative test results. Not false results, real results. Your error is in arriving at the number of tests that should have been conducted:

    That means that we may be behind in testing by 464,800 tests
    False. That means that we may be behind in testing by 464,800 positive test results.


    To find the number of tests that we are behind, we need the rate of positive to negative tests, then negative test results must be accounted in the total number of additional tests required to have tested every case.

Page 1 of 3 123 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •