Coronavirus prep

1…427428430432433…747

Replies

  • AnnPT77
    AnnPT77 Posts: 32,030 Member
    kshama2001 wrote: »
    AnnPT77 wrote: »
    kshama2001 wrote: »
    AnnPT77 wrote: »
    kshama2001 wrote: »
    @kimny72 and others following vaccine research closely - how do they get that 90%? Are they exposing the test subjects? That seems unethical, but if they are not, how would they know the subjects didn't get Covid because they simply were not exposed to it?

    https://www.statnews.com/2020/11/09/covid-19-vaccine-from-pfizer-and-biontech-is-strongly-effective-early-data-from-large-trial-indicate/

    ...The companies said an early analysis of the results showed that individuals who received two injections of the vaccine three weeks apart experienced more than 90% fewer cases of symptomatic Covid-19 than those who received a placebo.

    Once there are enough numbers, they decide it's an effect, not random chance. Statistics. (It's too long since I did any advanced stats, so I'm literally incapable of explaining the details, but it's a standard protocol.)

    I don't know enough about this study to know the control/treated matching protocols (how detailed), but the answer is still statistics.

    There are some challenge studies (where people are intentionally exposed) but I don't personally know whether Pfizer did any. It's not unethical to expose test subjects who sign up for the study with full knowledge (informed consent) that they are going to be intentionally exposed to the disease. Volunteers are solicited, and brave, generous people sign up for it, in the hope of speeding vaccine availability . . . if the vaccine they're helping test pans out.
    The_Enginerd wrote: »
    From the information I've read, an independent assessor checks progress at a set number of cases in the group. From there, they can compare the number of cases in those individuals given the placebo vs. the vaccine. It seems they are actually reaching the goals earlier than expected because the spread of the coronavirus has been so widespread, and they are already over halfway through the 164 total cases targeted for final review.

    Sources:
    https://www.vox.com/21556380/pfizer-biontech-covid-19-coronavirus-vaccine-optimism-skepticism
    https://blogs.sciencemag.org/pipeline/archives/2020/11/09/vaccine-efficacy-data

    Hmm, well 100% of the people I know IRL in Massachusetts who practice scrupulous mask wearing, hand washing, and social distancing have no symptoms of the virus, so if the volunteers are doing the same, how do we know the 90% vaccine effective rate has anything to do with the vaccine?

    While it is statistically unlikely that the vaccine group is better at taking precautions than the placebo group, it seems to be a stretch to award all 90% to the vaccine if the subjects were also taking precautions.

    The Vox article and another link within it do caution that the announcement comes from a press release, not validated data, so I will reserve judgement until more is known.

    FWIW, another statement I can't justify with details because my stats classes were so long ago: The articles I've read about the 90% thing don't sound like the reporters really understand the statistics, and are using some key numbers from a press release and not interpreting them insightfully. (I'm sure there are better technical/statistical analyses out there, but I'm not interested enough to seek them out, and I'd probably have to haul out textbooks or something to truly "get it" at the level I'd want, if I did bother. Don't wanna. I took a lot of stats back in the day, but I didn't enjoy it.)

    So, this is from memory: Your observation is that behaviors can have differed, leading to an accidental difference in cases among the two groups: Yes, absolutely. As the numbers of cases (considered in light of the test population size) go up, the probability that the difference comes from random factors like that goes down (in the technical sense of probability). Size of the test population matters hugely, too, in how probable it is that the test subjects reflect the population.

    Kind of think of it this way: If you flip fair coins, you get heads or tails, and over many tosses, you'll get about equal numbers of each, distributed in random patterns. (Not saying that's how viruses behave: Bear with me.) In any arbitrarily long sequence of coin flips, there is a non-zero probability that ALL flips will come up heads. But the more coin flips you do, that indefinitely long (theoretically possible) sequence of all-heads becomes less and less probable to occur as you go along. At some number of flips, the intelligent person will conclude it's overwhelmingly probable that the coin is rigged, even though a very, very long sequence of all-heads is possible.

    The analogous kind of thing can happen in a more complicated test (like the analysis of the control/test populations in a virus study). The math is much more complicated, as is the underlying real-world situation, but at a certain point, the numbers can reach a point where statistics tell you pretty clearly that it's extremely unlikely that the numbers you're getting are just from sheer chance, so the overwhelming probability is that the vaccine is actually working. There's still a chance (statistically) that your conclusion is wrong, but it's very tiny.

    This is a very structured statistical decision process, IMU, not just "lotta cases here, not so many there, I think we're good". Presumably, there will've also been opportunities to compare disease incidence in both arms to disease incidence in overall populations; or, to the extent there are decent-sized cohorts of the test population in a given locale, to compare results in both arms to the population incidence. While I don't believe that would be a core method of analysis, I'd be surprised if no one was looking at other hints like that (which would be less statistically valid methods by far, but could be sort of a gut-check).

    I did enjoy statistics and do remember the coin flipping :)

    However, with coin flipping, there are just two variables - heads and tails.

    With the Pfizer study, in addition to vaccine and placebo, there is also mask usage, hand washing, and social distancing. I'm curious as to how they control for taking precautions so that the 90% efficacy doesn't include that.

    The short and imprecise answer is that when you have enough people, the probabilities are that those kinds of factors average out - that the probability is that each group will have some hygiene-compulsive people, and some careless people, in roughly similar proportions to the populations from which they're drawn.

    At a certain point (numerically/statistically), the probability becomes very low that cautious people are overrepresented in the vaccine subgroup (vs placebo subgroup). (It remains possible the groups are unbalanced, but it's very improbable). It's still analogous (I chose that word carefully) to the coin-flips, even though it's not a binary or simple like the coin-flips.
  • Dnarules
    Dnarules Posts: 2,081 Member
    edited November 2020
    MikePfirrman wrote: »
    @ReenieHJ -- I hope your husband is OK.

    We should start talking more about IF you get it. I'm reading about oxygen sensors being used and the level at which you should go to the hospital.

    https://www.consumerreports.org/medical-symptoms/covid-19-pulse-oximeters-oxygen-levels-faq/

    Amazon has sensors as low as $17. Any other insight from people that know more than I do about science?

    On This Week in Virology, the infectious disease doctor mentioned monitoring oxygen stats, especially in the second week. We have one from Amazon. I made my sister in FL get one too.

    ETA. I see this was already mentioned by one of the other TWiV listeners :).
  • MikePfirrman
    MikePfirrman Posts: 3,307 Member
    Saw this for anyone interested. I'm into herbal therapies a lot.

    This is a uncommon Chinese herbal formula (at least in the Western culture), but when you look up the active ingredients, they are Quercetin (something we've mentioned a lot in here) and Resveratrol.

    http://www.china.org.cn/world/2020-11/09/content_76890331.htm

    https://www.sciencedirect.com/science/article/pii/S0753332219352746

    Now, herbal therapies can interact with prescription medications, so check with your docs. If I do end up getting it, I've got the everything but the kitchen sink to throw at it, LOL. I might throw that at it too.
  • lemurcat2
    lemurcat2 Posts: 7,885 Member
    kimny72 wrote: »
    I got one off Amazon, picked one that was a step up from the cheapest one jic :smile:

    I got one too, a while ago (I also got a thermometer). Been using the thermometer, but haven't tried the other yet -- this is motivating me to do so.
  • T1DCarnivoreRunner
    T1DCarnivoreRunner Posts: 11,502 Member
    edited November 2020
    Toilet paper hoarding is apparently happening again at Kroger. There aren't shortages, but the store is implementing limits to prevent the hoarding they are observing.

    Edit: Link https://www.10tv.com/article/news/local/kroger-limiting-purchases-on-paper-towels-toilet-paper-to-prevent-hoarding/530-9a27b476-b91f-4e86-a7e1-b14ac2357a97
  • mockchoc
    mockchoc Posts: 6,573 Member
    Chef_Barbell wrote: »
    Chef_Barbell wrote: »
    NY just ordered all gyms and restaurants closed at 10pm on Friday. No more take out liquor. I'm having a sad.

    So update on that... the governor meant a 10 o clock curfew on gyms and restaurants. Makes no sense since most gyms already close by that time and restaurants have to close by midnight.

    Our state numbers are up to 2.9% positive. Arbitrary rules make no sense to me.

    Does he not realise the virus is active 24 hours a day? Unbelievable. Also if they stop take out liquor some that are desperate will start smashing there way into the shops to get some.

Categories