Cigarbutt Posted October 1, 2020 Posted October 1, 2020 Italy is an interesting case. Herd immunity, apart from some regions and sub-regions where antibody prevalence reached relatively high levels, does not materially apply as most regions elsewhere showed relatively low exposure. Their initial outbreaks were quite traumatic (social mood) and behavioral changes happened in a big way, even given the cultural inclination for close contacts and multi-generational proximity. https://www.msn.com/en-ca/news/canada/1st-in-europe-to-be-devastated-by-covid-19-italy-redoubled-its-efforts-and-they-re-now-paying-off/ar-BB19Bca7?ocid=msedgntp TL;DR: A lot remains to be explained but their extensive and community-based test and contact tracing scheme, with a backward component, may have played an important role, when combined with other measures (that may have been too extreme, in retrospect). https://science.sciencemag.org/content/early/2020/09/29/science.abd7672 TL;DR: About 8% of confirmed cases explain about 2/3 of subsequent cases and 71% of confirmed cases did not spread. A variation of the Pareto principle at work. In terms of 'spread' analysis, some may be interested in: https://en.wikipedia.org/wiki/The_Tipping_Point https://www.theatlantic.com/health/archive/2020/09/k-overlooked-variable-driving-pandemic/616548/ TL;DR: There remain many unanswered questions but the cluster effect may not have been given enough consideration. The models used for Covid have been mostly deterministic, which makes sense for an overall assessment and when community spread reaches high levels (Covid and the flu behave similarly then, in terms of contagiousness and transmission). There is room for stochastic models. For those interested in this and in how insurance reserves are determined, there are deterministic models being used and these models have been improved (ie Bornhuetter–Fergusson Technique) by including a Bayesian component but, for some lines of business, deterministic models don't work very well (latent claims etc) and stochastic models can be incorporated to improve the visualization of potential outcomes (and for a virus, to define and apply cost-effective strategies to contain spread). Keeping community levels low and backward tracing can be self-feeding. An interesting aspect is that the stochastic and cluster-based approaches help to explain how good and bad luck may have a compounding effect, help to explain what Sweden, relatively, did right and did poorly and how Japan sort of got it right, given their historical path-dependent circumstances.
cwericb Posted October 2, 2020 Posted October 2, 2020 Trump's right hand, Hope Hicks has Covid-19. Showed symptoms last night and has been in close contact with Trump & staff. Trump waiting for test results. May have to quarantine.
Dalal.Holdings Posted October 2, 2020 Posted October 2, 2020 Trumpf and Melania positive... 2020 still doing 2020 You literally couldn’t make this up: Mr. Trump, who for months has played down the seriousness of the virus and hours earlier on Thursday night told an audience that “the end of the pandemic is in sight,” will quarantine in the White House for an unspecified period of time, forcing him to withdraw at least temporarily from the campaign trail only 32 days before the election on Nov. 3
LC Posted October 3, 2020 Posted October 3, 2020 So here we are at the start of October, it may be useful to revisit our good ole' excess mortality data. Europe: https://www.euromomo.eu/graphs-and-maps USA: https://www.cdc.gov/nchs/nvss/vsrr/covid19/excess_deaths.htm Looks like most of Europe has stabilized, with the exception of Spain. USA looks like it has fully passed the second wave and mortality has normalized. So the big question is whether there will be a third wave or whether COVID-19 has run its course...
Cigarbutt Posted October 4, 2020 Posted October 4, 2020 ^Yes, excess mortality has been useful for analysis and it contributes now to a gradually clearer picture, helpful for constructive 'blame'..(assuming limiting mortality in a cost-effective manner is the objective). For disclosure, my area (province), using this benchmark, has done very poorly (close to Spain). The problem around here is not really ideological but more competence-related (previous decisions and those taken in the heat of the moment) but it's a real head-scratcher to see that the most prevalent mode of attitude is "it is what it is", which makes me wonder if my area is ready for more collective challenges (debt, socialized risk etc). Anyways, the following is interesting when evaluating the value of excess mortality data. Yesterday, in the US, the percent positive rate was at 5.7% (5.7% in October 2020!, no wonder the virus is not only trickling down), so you can expect (like it has happened recurrently along the virus spread evolution) reporting lags to cause the updated excess recent mortality curve to be adjusted upwards for some time because there continues to be a low grade excess mortality going on (in October 2020!) and it looks like this will carry on for a while. Interestingly, from many sources, it looks like the average years lost per mortality is between 10 and 15 years. https://weinbergerlab.github.io/excess_pi_covid/ You also may be interested in the following. Last May, they looked at some aspects of the meaning of this data and, recently, have looked at data more globally. https://voxeu.org/article/excess-mortality-england-european-outlier-covid-19-pandemic https://voxeu.org/article/us-excess-mortality-rate-covid-19-substantially-worse-europe-s An interesting way to look at this: Opinion: There is so much to learn and yet so little appetite to do so where it is most needed (both backward and forward looking).
Gregmal Posted October 4, 2020 Posted October 4, 2020 And here we go again, loser De Blasio trying to start locking down NYC again lol.
Investor20 Posted October 5, 2020 Posted October 5, 2020 https://www1.nyc.gov/site/doh/covid/covid-19-data-testing.page New York City Percent positive for antibodies You can scroll down in above page to "Antibody Tests" click on percent of people tested positive. I couldn't get the image here, so downloaded the data associated with that plot using "get data" and given in below table. It started in April with 0.613, that is 61.3% seroprevalence in NYC! And ofcourse people got infected afterwards and there are reports of antibodies lasting only for 2 months, at least in part of the population. That is total infected in NYC is more than 61.3%. Am I wrong? Why so many report it is 20% infected in NYC? weekdate Percent positive 4/11/2020 0.613 4/18/2020 0.439 4/25/2020 0.328 5/2/2020 0.323 5/9/2020 0.343 5/16/2020 0.344 5/23/2020 0.33 5/30/2020 0.298 6/6/2020 0.281 6/13/2020 0.254 6/20/2020 0.233 6/27/2020 0.215 7/4/2020 0.201 7/11/2020 0.197 7/18/2020 0.204 7/25/2020 0.208 8/1/2020 0.212 8/8/2020 0.206 8/15/2020 0.194 8/22/2020 0.201 8/29/2020 0.201 9/5/2020 0.202 9/12/2020 0.197 9/19/2020 0.202 9/26/2020 0.204
Spekulatius Posted October 5, 2020 Posted October 5, 2020 ^ I have Seen a lot of questionable analysis in their thread, but this is beating almost all of it. Adding up all the percent positives to get the total cumulative positives would only work if everyone in NY would be tested every week and everyone tested positive once eliminated from the pool of subsequent testing. Of course that’s not the case, the number tested each week is only a tiny part of the total population.
Investor20 Posted October 5, 2020 Posted October 5, 2020 ^ I have Seen a lot of questionable analysis in their thread, but this is beating almost all of it. Adding up all the percent positives to get the total cumulative positives would only work if everyone in NY would be tested every week and everyone tested positive once eliminated from the pool of subsequent testing. Of course that’s not the case, the number tested each week is only a tiny part of the total population. Its not cumulative. For cumulative the numbers will increase with time. The seroprevalence as per NYC website went down from 61.3% in April to 20% in September. That indicates people loosing antibodies, which I have posted before as a problem at looking at seroprevalence to check how many got infected. If you do cumulative it will be 61.3% by April plus however many infected after April. The numbers in the data base are given as fractions 0.613, is 61.3%. Look at the plot on the website. I also attached a screenshot of the plot with this message.
Cigarbutt Posted October 5, 2020 Posted October 5, 2020 @Investor20 This is basic statistics. If you want to generalize the sample to population, there are basic rules to follow. The antibody testing that is being reported is using serial samples for which inclusion criteria did not tightly apply to the general population and for which rules (mostly self-selection) changed over time.. Look at the 'public' side of the question: https://www1.nyc.gov/site/coronavirus/get-tested/antibody-testing.page Quite frankly, the "strategy" behind the antibody testing here is perplexing. It's not as if the area doesn't have other challenges to address. If you look at the numbers in and around NYC, herd immunity is playing a peripheral role although it seems that they (collectively, as a result of individual and non-coalesced decisions) use an approach that will guarantee success, if herd immunity is the way you define success.. Even if the antibody data is crude and poorly informative, the picture suggests that that when the virus will have run out of easy targets in the Bronk, Queens etc, it will materially make its way into Manhattan. What will happen in the NYC area is the key question and the picture will continue to be dynamic (and the key variables other than herd immunity are pointing to better outcomes) but the exercise is much more mathematical than ideological.
Investor20 Posted October 5, 2020 Posted October 5, 2020 The NYC website is reporting 61.3% seroprevalence for NYC as of 4/11/2020, not for a borough or an area. Cigabutt, are you saying NYC is reporting wrong numbers? https://gothamist.com/news/nyc-map-coronavirus-antibody-testing-data-date NYC Releases Largest And Most Detailed Coronavirus Antibody Testing Data To Date, Aug 20, 2020 "At one point, the positivity rate of antibody testing soared to as high as 61% in early April. It has since leveled off to around 20% over the last three weeks." A lot of people seem to loose antibodies in very short time. The 61% on April 11th went to 34% on May 9th. Thats very concerning from vaccine point of view.
KCLarkin Posted October 5, 2020 Posted October 5, 2020 The NYC website is reporting 61.3% seroprevalence for NYC as of 4/11/2020, not for a borough or an area. Not sure why you have such a hard time with stats. The data for 4/11 is a tiny sample and is an obvious outlier. The question is why?
Cigarbutt Posted October 5, 2020 Posted October 5, 2020 ^Are you doing this on purpose 'cause it feels like fighting a barbarian invasion with a rot a the core? On April 11th, they reported a 61.3% antibody+ rate on the tests they performed. On April 11th, they performed 1131 tests, which is about 0.01% of the NYC population. On April 11th, i spoke to people on the field and the last thing they cared were irrelevant data points. On April 11th, NY reported 783 additional deaths of COVID-19 (daily count), bringing the death toll to 8,627. On April 11th, NY reported total Covid-19 cases increased by 9,946, to a total of 180,458. That day, the national death toll was at 18,578. You've got to focus on the relevant numbers.
Investor20 Posted October 5, 2020 Posted October 5, 2020 The NYC website is reporting 61.3% seroprevalence for NYC as of 4/11/2020, not for a borough or an area. Not sure why you have such a hard time with stats. The data for 4/11 is a tiny sample and is an obvious outlier. The question is why? Sample size of 1131 is not that small. Its typically used for national polls. https://www.scientificamerican.com/article/howcan-a-poll-of-only-100/ How can a poll of only 1,004 Americans represent 260 million people with only a 3 percent margin of error?
KCLarkin Posted October 5, 2020 Posted October 5, 2020 Sample size of 1131 is not that small. Its typically used for national polls. https://www.scientificamerican.com/article/howcan-a-poll-of-only-100/ How can a poll of only 1,004 Americans represent 260 million people with only a 3 percent margin of error? My guess is this sampling bias or a data error. Regardless, it is clearly an outlier.
Peregrine Posted October 6, 2020 Posted October 6, 2020 One thing we've learned about antibodies is that many who had mild symptoms or were completely asymptomatic didn't develop detectable antibodies and that antibodies wane over time. I think NYC's 20-25% seropositivity level is most likely an underestimate of the true number of infected - and possibly by a lot.
muscleman Posted October 6, 2020 Posted October 6, 2020 One thing we've learned about antibodies is that many who had mild symptoms or were completely asymptomatic didn't develop detectable antibodies and that antibodies wane over time. I think NYC's 20-25% seropositivity level is most likely an underestimate of the true number of infected - and possibly by a lot. That's what I kept saying for a long time. We already have herd immunity in NYC and a few other places. But people here who desperately want to see Trump failing and therefore having strong confirmation bias kept pointing to the 80% infection rate for achieving herd immunity.
KCLarkin Posted October 6, 2020 Posted October 6, 2020 That's what I kept saying for a long time. We already have herd immunity in NYC and a few other places. Yes, you've been wrong about this for a long time. I know you've sorta apologized for this. But you predicted 20k daily cases by the end of August. It is October and the U.S. is still over 40k cases. Cases are up since your mea culpa. Now you are using NYC as proof of herd immunity when they just partially re-opened indoor dining? And already have early signs of a second wave?
Peregrine Posted October 6, 2020 Posted October 6, 2020 Now you are using NYC as proof of herd immunity when they just partially re-opened indoor dining? And already have early signs of a second wave? NYC's positivity rate: Hardly signs of a second wave.
KCLarkin Posted October 6, 2020 Posted October 6, 2020 Hardly signs of a second wave. Doubt Cuomo will allow a full second wave. But increasing positivity and Rt > 1 despite heavy restrictions are clear indicators that herd immunity threshold hasn’t been reached. Edit to add: schools and restaurants haven’t even been open a full incubation period, so too early to detect a second wave.
Peregrine Posted October 6, 2020 Posted October 6, 2020 Hardly signs of a second wave. Doubt Cuomo will allow a full second wave. But increasing positivity and Rt > 1 despite heavy restrictions are clear indicators that herd immunity threshold hasn’t been reached. Edit to add: schools and restaurants haven’t even been open a full incubation period, so too early to detect a second wave. There's a big misconception about what the herd immunity threshold even is. For many, it seems to mean the point at which 1) the virus disappears or 2) that the virus no longer spreads. But that's not possible given the thousands of different respiratory viruses that circulate normally. Rather, the herd immunity threshold is better described as the point in which the virus crosses over from epidemic to endemic. Under that definition, HIT being reached and Rt>1 is not incompatible. The virus will still be around long after the epidemic has ended. So it's because of Cuomo's excellent leadership that a second wave will be avoided? Florida's rt has been estimated as below 1 since late June, even below that of NY's...by your thinking, was it DeSantis's doing as well?
orthopa Posted October 6, 2020 Posted October 6, 2020 Couple things about antibodies. The antibody testing needs to be delineated between IgM and IgG to get a sense of timing and prevalence. (my assumption is that is what was being tested as that's what was commonly being tested across the country) Both become detectable and and wane at different time periods. Over time "months" nearly all who were infected will have their antibodies wane down to not detectible. Many getting tested in August could have "had antibodies" in March, April, May etc but show up negative in August. FWIW earlier readings have the likelihood as being most accurate as the start date for large scale exposure and initial community spread (and thus catching the start of the antibody curve if you will) is much closer to March then now. I haven't looked through the data being discussed but if a representative population was ~62% positive in early March that is a much more accurate level then what you would get now IMO. Consider that a floor, if again representative. Also clearly explains NYC infection curve (very low level of cases for months and months) AND their over representative death count and spike in spring. Losing antibodies as delineated above is not uncommon and does not immediately point to reinfection or possibility. Look at worldwide re infection examples. Handful out of millions of cases. T cell immunity is what you guys are missing here. All the info you guys are looking for in NYC is in T cell immunity. The problem is there is no way do detect/test for such at this time. I attached a graph giving a rough idea of what I am trying to convey. As it says chart for illustrative purposes only.
orthopa Posted October 6, 2020 Posted October 6, 2020 Hardly signs of a second wave. Doubt Cuomo will allow a full second wave. But increasing positivity and Rt > 1 despite heavy restrictions are clear indicators that herd immunity threshold hasn’t been reached. Edit to add: schools and restaurants haven’t even been open a full incubation period, so too early to detect a second wave. There's a big misconception about what the herd immunity threshold even is. For many, it seems to mean the point at which 1) the virus disappears or 2) that the virus no longer spreads. But that's not possible given the thousands of different respiratory viruses that circulate normally. Rather, the herd immunity threshold is better described as the point in which the virus crosses over from epidemic to endemic. Under that definition, HIT being reached and Rt>1 is not incompatible. The virus will still be around long after the epidemic has ended. So it's because of Cuomo's excellent leadership that a second wave will be avoided? Florida's rt has been estimated as below 1 since late June, even below that of NY's...by your thinking, was it DeSantis's doing as well? I don't know why some people want it to be bad or get worse. So if NYC is not reaching or has reached the point of "herd immunity" what explains the positivity rate in the graphic you posted? Mask wearing? Social distancing? I live in NY. Believe me, everyone isn't wearing masks all the time or social distancing all the time and many, many restaurants are packed and busy on a nightly basis where you don't have to wear a mask if you eating. Sure NY is better then TX and FL at social measures but looking at frank87's graphic there is a larger force at play, one that doesn't slip up like humans do.
Investor20 Posted October 6, 2020 Posted October 6, 2020 Couple things about antibodies. The antibody testing needs to be delineated between IgM and IgG to get a sense of timing and prevalence. (my assumption is that is what was being tested as that's what was commonly being tested across the country) Both become detectable and and wane at different time periods. Over time "months" nearly all who were infected will have their antibodies wane down to not detectible. Many getting tested in August could have "had antibodies" in March, April, May etc but show up negative in August. FWIW earlier readings have the likelihood as being most accurate as the start date for large scale exposure and initial community spread (and thus catching the start of the antibody curve if you will) is much closer to March then now. I haven't looked through the data being discussed but if a representative population was ~62% positive in early March that is a much more accurate level then what you would get now IMO. Consider that a floor, if again representative. Also clearly explains NYC infection curve (very low level of cases for months and months) AND their over representative death count and spike in spring. Losing antibodies as delineated above is not uncommon and does not immediately point to reinfection or possibility. Look at worldwide re infection examples. Handful out of millions of cases. T cell immunity is what you guys are missing here. All the info you guys are looking for in NYC is in T cell immunity. The problem is there is no way do detect/test for such at this time. I attached a graph giving a rough idea of what I am trying to convey. As it says chart for illustrative purposes only. Agree there is clear waning of antibodies in not only NYC data but observed many places elsewhere. I am reading the 20% present new york antibody as simply last several months of infections, not cumulative infections from February. My understanding of T Cell immunity is it does not stop infection but reduces severity of infection. With this regard vaccines might not prevent re-infection, but only severity of infection: “if everyone gets vaccinated and we continue to implement the public-health measures that I have been talking about incessantly over the last several months,” Fauci said according to below article https://www.self.com/story/fauci-masks-after-vaccine
clutch Posted October 6, 2020 Posted October 6, 2020 I don't know why some people want it to be bad or get worse. Obviously, so that they can keep blaming Trump... ::)
Recommended Posts