Chickenboy
Posts: 24520
Joined: 6/29/2002 From: San Antonio, TX Status: offline
|
quote:
ORIGINAL: MakeeLearn quote:
ORIGINAL: MakeeLearn Cuomo Says 21 Percent of Those Tested in N.Y.C. Had Virus Antibodies Apr 23, 2020 https://www.nytimes.com/2020/04/23/nyregion/coronavirus-new-york-update.html "About 21 percent of roughly 1,300 people in New York City who were tested for coronavirus antibodies this week tested positive, Gov. Andrew M. Cuomo said on Thursday. The results come from a state program that randomly tested 3,000 supermarket customers across New York State. Nearly 14 percent of those tests came back positive, Mr. Cuomo said. If those numbers translate to the true incidence of coronavirus, they would mean that more than 1.7 million people in New York City, and more than 2.6 million people statewide, have already been infected. These numbers are far greater than the 250,000 confirmed cases of the virus itself that the state has recorded. They would also mean that the fatality rate from the virus is relatively low — about 0.5 percent, Mr. Cuomo said." No comments??????????????????? This goes along with the studies coming out of California. I followed the article and have some comments as I see them: Here's a more detailed discussion of the serology testing than Cuomo's briefing summary: https://www.nytimes.com/2020/04/23/nyregion/coronavirus-antibodies-test-ny.html First off, the testing was not a random subset of the population of either the City of New York or the State of New York. It was voluntary testing of 'shoppers'. As stated in some observations of the results, it did not include children, teenagers and older adults that may have been sheltering in place. Supermarket shoppers are probably not a representative sample when a state is in lockdown. Sampling 3,000 people out of a 14M population is a sample rate of .002% of the state population. Statistically significant? Not even close. Not even with a test that is 100% sensitive and 100% specific-which this serology test is most certainly not. It gets even worse when you start talking about interpretation of some of the commercial 'quicky' serology tests that will soon be flooding the market. None of the relevant health histories were taken from the shoppers. How many of these people had a FUO (Fever of Unknown Origin) or other clinical signs going back to-say-December 1, 2019? Dunno. So now you lost your 'asymptomatic' versus 'symptomatic' distinction. Were any of the shoppers known positive by RT-PCR (and recovered)? Dunno. Now you lost your correlation between RT-PCR (virus positive) and convalescent serum response too. The tests, while conducted in a respected State laboratory, were intentionally interpreted to err on the side of caution. Namely what constitutes a 'positive' finding for the test. I couldn't find details on what the test parameter cutoffs were for calibration purposes, but this likely understates the number of people with serologic antibody titers-they may have been below the cutoff in their profile. There appear to be disparate results between NYC samples and 'outstate' samples: 21% of NYC, 17% in Long Island, 12% in Westchester / Rockland and less than 4% in the rest of the state. There was also an apt quote from one of the public health folks that cautioned against using antibody testing as a criterion to allow people into the world, "To use it as a fulcrum for when someone can travel or work...we have to be extremely careful as a society in doing that." Little is known about the meaning of being seropositive with respect to immunity or resistance to future infections. It will be extremely problematic to try to apply these disparate findings to any sort of statewide 'back to work' edict. If only 21% of New Yorkers have been 'exposed' to the virus, is that sufficient for herd immunity? No. You probably need ~70% for that. So NYC is less than a third of the way 'there'. Anybody here think NYC can stay closed for another 6 months? Plus, shelter-in-place edicts will slow the community spread of the virus, making it tougher to 'hit the numbers' for herd immunity. If the target is 70% before herd immunity is established, NYC may *never* get there with their extreme quarantine. And what about the rest of the state? 4% seroconversion outside of NYC? That's nothing. How can the governor of the state apply local observations about NYC to the greater population with less than 1/5 of the seroconversion? Keep the state closed down until the outstate seroconversion rate is up to 70%? So, to summarize: You have non-random sampling of residents of your state. You get different results depending on geography. Not a surprise there, as your virus testing and case/mortality data already noted the glaring difference between NYC and everywhere else. Nothing about the data suggests that you are anywhere close to 'herd immunity'-on a novel pathogen such as this-one doesn't know what herd immunity numbers will even look like anyways. Per the quote above, you can't use this data as a decision making tool regarding a return to travel or work or some 'immunity passport'. Earlier in the thread I cautioned about the overinterpretation of the meaning of serology results. This study highlights some of the fundamental problems using this retrospective data (serologic antibodies are always retrospective data re: exposures) to make meaningful decisions moving forward. Without proper stratification of your sample and profiling samples relative to clinical history you cannot derive meaning from the study for prospective decisions. Farther still to using this data for a disparate and heterogenous state population. IFR is lower than initially thought? Not a surprise. That happens in lots of diseases (sometimes it's higher too). Big whoop. Other than an "I told you so" gotcha moment, how does this help making decisions moving forward beyond what you already know? So is it all garbage? No. But without some of the needed study design and implementation, it's impossible to meaningfully interpret or use as a decision making tool. Therefore these studies will be rendered a scientific curiosity or a 'gee whiz'. We will have to move forward with incomplete information. The 'numbers' won't obviate our need to make tough and balanced decisions. And we have to be on guard for overreliance on these sorts of tests to be some sort of sop for the masses. Those are my thoughts.
< Message edited by Chickenboy -- 4/24/2020 4:32:35 PM >
_____________________________
|