Do you need to vent about things COVID-19?

[Madelaine McMasters]

2 hours ago, Qie Niangao said:

5 hours ago, AyelaNewLife said:

• Daily death tolls have one use, and one use only - to monitor trends. That becomes useless if you mix up the methodology part-way through this outbreak, especially if you intend to add highly unreliable and varied data with a lag time of between hours and days into the mix. So yes, anyone calling for care home deaths to be added to the daily hospital deaths total is an idiot. Total suspected deaths has a use as a separate measure, and will by its very nature require a lag time of probably a week and looking at week-long snapshots in order to be useful.

Wow, maybe there's some local reason that deaths in long term care facilities are different from those in hospitals, but here in Canada it would be impossible to "monitor trends" without counting those deaths -- because they are the trend!  Early-on, incidents of CV19 were pretty much exclusively among international travellers, then those in close contact with travellers, then broader contacts and community spread. And then it started entering long-term care homes, and exploding. To somehow exclude those deaths on the grounds of methodological purity would be to miss a huge expansion in the deadliness of this disease -- which seems only to artificially minimize the pandemic.

Maybe that's not what's being suggested.

What I think Ayela is arguing is not that nursing home deaths be ignored, but that they not be introduced into the current trend data if they weren't already there. Regardless how broken a particular methodology is, changing it mid stream makes it more difficult to get at the underlying trend. Every measurement we currently have is crap, but so long as those measurements are consistent crap, they may correlate to actual viral spread. If Canada has been counting nursing home deaths all along, they should continue to do so. If they haven't, then including them now would also require doing some sort of curve shifting for all the previous data. Not doing this would insert a discontinuity into the trend line that could compromise any ability to assess containment progress.

All that said, some of the measurements correlate better to other time varying things, like test or facility capacity. You can't have more certified cases than your ability to certify them via tests. You can't hospitalize more patients than your system can accept. Death count is probably the most rugged statistic, as it's the smallest of the tracking numbers, the least limited by measurement ability, and the least subject to difference of interpretation. By the time someone dies of COVID-19, the practitioners are fairly certain, via testing, symptomatology or both, that the death was caused by COVID-19.

[Selene Gregoire]

Longstanding issues put Native American communities at high COVID-19 risk

[Kitteh Whiskers]

The whole quarantine thing is getting me grumpy .Well hungry and grumpy lol. The officials say that here in L.A. the lock down will be lifted May 15th but, I have a hard time believing that. People here are still going out and partying 😐

 

 

[Scylla Rhiadra]

2 hours ago, Madelaine McMasters said:

What I think Ayela is arguing is not that nursing home deaths be ignored, but that they not be introduced into the current trend data if they weren't already there. Regardless how broken a particular methodology is, changing it mid stream makes it more difficult to get at the underlying trend. Every measurement we currently have is crap, but so long as those measurements are consistent crap, they may correlate to actual viral spread. If Canada has been counting nursing home deaths all along, they should continue to do so. If they haven't, then including them now would also require doing some sort of curve shifting for all the previous data. Not doing this would insert a discontinuity into the trend line that could compromise any ability to assess containment progress.

I totally understand the point that you are making, but . . . this isn't a lab experiment, and it's not happening under controlled conditions.

Nor is the point of data to extrapolate a conclusion or hypothesis: it's about applying it to dynamic models, the purpose of which is to help us produce effective responses in real time. And you can't do that if you decide that the "purity" or scientific validity of the data is more important than understanding how the really messy situation is unfolding on the ground.

I think it may depend mostly upon how we are contextualizing the data and the models that result from its application. If we are not highlighting the fact that, for instance, we are now including deaths that did not occur in a hospital, the model looks skewed -- or more precisely, it is prone to be misread and misunderstood. So, at various times, China, France, and the US have all seen massive spikes in deaths or new cases that are, in fact, the result of applying new criteria to how we amass the data. To someone who doesn't understand what has happened, then, yeah, that is going to misinform. So, you need to make sure that they are informed. As for scientists, epidemiologists, and so forth, one trusts that they are sophisticated enough in their understanding of statistics to get the true significance (or lack thereof) of such spikes, and to account for that.

My take on what Ayela (and Lil) are getting at is that sudden shifts in the data due to a change in criteria can, and in some cases are being used for essentially political purposes. And that's less the fault of the data, or how it is being collected, and more about how journalists and politicians may be intentionally (or, to be fair, perhaps out of ignorance too) misreading, and misrepresenting the data.

Which is where we come to this:

5 hours ago, Qie Niangao said:

Wow, maybe there's some local reason that deaths in long term care facilities are different from those in hospitals, but here in Canada it would be impossible to "monitor trends" without counting those deaths -- because they are the trend!  Early-on, incidents of CV19 were pretty much exclusively among international travellers, then those in close contact with travellers, then broader contacts and community spread. And then it started entering long-term care homes, and exploding. To somehow exclude those deaths on the grounds of methodological purity would be to miss a huge expansion in the deadliness of this disease -- which seems only to artificially minimize the pandemic.

The point that Qie makes in the Canadian context underlines, to my mind, the importance of keeping models flexible and dynamic, and responsive to real world situations. Canada has always included nursing home deaths in the count, but that data has not always been added in a completely timely way.

But even if it was entirely new data, it would still be important because, as Qie notes, this is where the battle is being fought right now. The last two weeks have seen Canada's infections plateau, and the rate of growth of infections drop steadily. We seem, at the moment, to have the contagion at least somewhat under control. Except in nursing homes: our mortality rate is not falling in tandem with infection rates because the virus is now tearing through nursing homes and killing dozens.

That is important, not merely in an ethical way, but in understanding how we need to shift our strategy, and reallocate resources. And it would be no less important, and maybe more so, if we hadn't been tracking nursing homes from the beginning.

We need to know this stuff -- politicians, journalists, scientists, and people on the street. The key is making sure that it is presented in such a way that is contextualized and properly interpreted.

[Madelaine McMasters]

3 minutes ago, Scylla Rhiadra said:

I totally understand the point that you are making, but . . . this isn't a lab experiment, and it's not happening under controlled conditions.

My take on what Ayela (and Lil) are getting at is that sudden shifts in the data due to a change in criteria can, and in some cases are being used for essentially political purposes. And that's less the fault of the data, or how it is being collected, and more about how journalists and politicians may be intentionally (or, to be fair, perhaps out of ignorance too) misreading, and misrepresenting the data.

Which is where we come to this:

The point that Qie makes in the Canadian context underlines, to my mind, the importance of keeping models flexible and dynamic, and responsive to real world situations. Canada has always included nursing home deaths in the count, but that data has not always been added in a completely timely way.

Are we discussing the models, or the publicly reported data?

There's a reason to exclude, or to handle differently, nursing home data if the goal is to judge the effectiveness of public containment efforts. Nursing homes are an atypical environment. They are effectively observing "stay at home", but with devastating results do to population density, atypical spread mechanisms,  and patient vulnerability. Tossing nursing home deaths into the general population data, without factoring in those differences could well result in counterproductive public policy. Analyzing nursing data apart from the general public might allow design of practices to minimize deaths in that specific population while minimizing the economic impact of prolonged severe lock down policies.

There's also a huge hole in the data set from deaths at home, of people unwilling or unable to seek or get treatment. It may take years to tease that out of overall death rates. I expect death counts to be increased retroactively in the months and years to come.

If the COVID models are sophisticated enough to handle nursing home data, I doubt they are looking at death statistics the way we do. As you've suggested, the publicly available statistics are necessary oversimplified and subject to politicization.

This all reminds me of the endless debates over climate science data. Every few years, adjustments are made to historical data sets as the result of discovering systemic errors or biases in the data. After significant deliberation, the community decides how to make those historical adjustments. Those adjustments are inevitably met with cries of foul play.

COVID-19 will be like that... and then some.

I think nursing homes are a different enough animal to warrant their own COVID-19 statistics, highlighting shortcomings in our elder care systems.

[Qie Niangao]

47 minutes ago, Madelaine McMasters said:

There's also a huge hole in the data set from deaths at home, of people unwilling or unable to seek or get treatment. It may take years to tease that out of overall death rates. I expect death counts to be increased retroactively in the months and years to come.

Yeah. There's another problem of potentially massive undercounting from the very beginning of the pandemic: deaths by non-respiratory symptoms of COVID-19. There's still a huge hole here, months-in, as we learn just how much of the virus' mortality is due to its effect on the circulatory and renal systems -- and those are just what's become obvious so far. This virus is not limited to its most apparent (and only probably most common) respiratory syndrome. 

Thing is, for that particular limitation in the data, I don't see a problem with keeping those deaths separate for trend analysis, because there's no reason to think their proportion of virus mortality is changing over time. Exactly the same for quibbles about how much of a death needs to be COVID-caused: I'm 100% fine with counting all deaths with positive tests as part of the pandemic because that's how it's always been done. (The valid concern here is that we see more of them as testing improves, but that's just a universal truth that colors all the statistics.)

But the same is simply not true of deaths in long-term care facilities. There's every reason to think they make up a larger proportion of deaths now than before. If we're looking at trends to decide what public health restrictions to lift or impose, those deaths must be included because they are all about how those restrictions are working or falling short at this stage of the pandemic. There's no way to make responsible policy decisions ignoring them.

[Madelaine McMasters]

14 minutes ago, Qie Niangao said:

But the same is simply not true of deaths in long-term care facilities. There's every reason to think they make up a larger proportion of deaths now than before. If we're looking at trends to decide what public health restrictions to lift or impose, those deaths must be included because they are all about how those restrictions are working or falling short at this stage of the pandemic. There's no way to make responsible policy decisions ignoring them.

Yeah, but include them separately, in their own column. I suspect the models weigh nursing home, and possibly other dense housing (prisons, etc) deaths a little differently, simply because you can't use the same mitigation techniques there. If you dump all the deaths together into one sum, your resulting policy is going to be pretty blunt.

Edited April 19, 2020 by Madelaine McMasters

[Qie Niangao]

20 hours ago, Seicher Rae said:

• Five of those states had protests against stay-at-home orders, and all of the photos show mobs of angry white people in MAGA hats, jammed together, no masks.

They were incited by Outside Agitators -- in the White House. And celebrated in right-wing media. And advisor to the president, Stephen Moore (the "economist", not Stephen Miller the reincarnation of Goebbels) recognized their contributions thus:

Quote

“I call these people the modern-day Rosa Parks — they are protesting against injustice and a loss of liberties,” Moore said of the protesters.

I mean...

[Qie Niangao]

21 minutes ago, Madelaine McMasters said:

Yeah, but include them separately, in their own column. I suspect the models weigh nursing home, and possibly other dense housing (prisons, etc) deaths a little differently, simply because you can't use the same mitigation techniques there. If you dump all the deaths together into one sum, your resulting policy is going to be pretty blunt.

It's certainly a factor that can be used in analysis, but there are also many other factors that may be relevant to specific policy decisions. I guess I'm arguing that if the associated deaths are tallied in a separate "column", that really needs to be a column in a notional "pivot table" of these factors. For some purposes, living arrangements such as long-term care residence is of interest, sometimes perhaps age, sometimes maybe past exposure to a specific different virus, proximity to chemical plants, etc., which may be (quasi-)independent contributors, each relevant to one policy or another.

[AyelaNewLife]

8 hours ago, Qie Niangao said:

Wow, maybe there's some local reason that deaths in long term care facilities are different from those in hospitals, but here in Canada it would be impossible to "monitor trends" without counting those deaths -- because they are the trend!  Early-on, incidents of CV19 were pretty much exclusively among international travellers, then those in close contact with travellers, then broader contacts and community spread. And then it started entering long-term care homes, and exploding. To somehow exclude those deaths on the grounds of methodological purity would be to miss a huge expansion in the deadliness of this disease -- which seems only to artificially minimize the pandemic.

Maybe that's not what's being suggested.

 

11 hours ago, AyelaNewLife said:

• Daily death tolls have one use, and one use only - to monitor trends. That becomes useless if you mix up the methodology part-way through this outbreak, especially if you intend to add highly unreliable and varied data with a lag time of between hours and days into the mix. So yes, anyone calling for care home deaths to be added to the daily hospital deaths total is an idiot. Total suspected deaths has a use as a separate measure, and will by its very nature require a lag time of probably a week and looking at week-long snapshots in order to be useful.

Emphasis mine.

In the UK, we have had a daily press conference delivered by a senior minister or two, and a senior scientific adviser or two. These briefs usually come with a few graphs, based around the key metric of hospital deaths with confirmation of Covid 19 each day - the press conference is at 5pm (4pm weekends), and the data covers the 24 hour period leading up to the previous 5pm.

We have a strong media campaign here in the UK from the ignorant and the disingenuous to include deaths from suspected Covid-19 from the care sector into this very specific daily measure. It has been explained almost daily for two weeks now in the Q&A part of these briefings why that's nonsensical, and why monitoring suspected deaths from outside hospitals is also key to their understanding of the outbreak, but that it must be done in a way appropriate to the information concerned.

The methodology chosen by the ONS (Office for National Statistics) is effectively a sift of all death certificates registered in a certain period. Death certificates should be registered within five days, but exceptions may apply when coroners need to conduct a longer investigation (relatively uncommon in care home Covid-19 cases). Most are however not registered within a day; and the standard amount of time it takes varies wildly due to regional variations in both the care sector and local authorities. In practice, that means in some cases you could have, say, 5% of suspected Covid-19 deaths registered in time to be included in the daily statistics the following day, while on another day you could have 40-50% of suspected deaths registered in time. This data would be chocolate teapot levels of useless.

However, if you leave it about a week prior to gathering said data, you can get a far more reliable set of data. It might only cover 80-90% of suspected deaths over that period, but the relative variation is far, far smaller and can (and has!) actually be used to monitor trends outside the hospital environment. And remember, unlike journalists our scientific advisers are not idiots; they will not be looking at a single graph and basing decisions off that. They're taking into account a great many metrics, each treated appropriately based on individual circumstances.

The idea that no one is paying attention to care home deaths is raw fantasy - this is a complaint manufactured newspapers more interested in their morbid scorecards, and tabloid readers peddling half-truths. I don't think anyone here is arguing that, however... but plenty of idiots out in "the real world" are, and it's those I'm ranting about.

Edited April 19, 2020 by AyelaNewLife

[Qie Niangao]

Ah, so the distinction is really between daily and weekly totals, and in the UK apparently the hospital data is of some value on a daily basis. The Canadian data (from all sources, AFAIK, and not only deaths either) are just garbage for up to a week -- due to reporting lag if nothing else -- so we've gotten pretty accustomed to reading the charts with blinders on the last few bars. And that is pretty frustrating when policy makers are eager to see whether interventions are succeeding.

[Seicher Rae]

Because apparently it is impossible to get sick at a beach...

for the full article: https://www.washingtonpost.com/nation/2020/04/18/florida-beaches-reopen-coronavirus/

[Luna Bliss]

5 hours ago, AyelaNewLife said:

The idea that no one is paying attention to care home deaths is raw fantasy - this is a complaint manufactured newspapers more interested in their morbid scorecards, and tabloid readers peddling half-truths. I don't think anyone here is arguing that, however... but plenty of idiots out in "the real world" are, and it's those I'm ranting about.

In the US too much time elapsed before we knew of the dire circumstances affecting a great proportion of nursing homes. Once this was more known of course the reporters would come out of the woodwork looking to blame someone and get those headlines. It's annoying that some misinterpret statistics in their blaming efforts, but having science people to interpret better would not have solved the main problem, which lies deeper than the interpretation of statistics.

There was a disconnect between nursing homes and the rest of society, one that would have been alleviated with a pandemic preparedness plan where better communication could have taken place.  If we had known so many were dying in nursing homes then some families would have been able to get grandma out  before lockdown occurred.

So the question for me is, why didn't we have a pandemic preparedness plan? Apparently we did have one (I'd like to see if it addressed homes where large populations are housed), but it was tossed by Trump a year or more ago, and other programs that could have helped reduce deaths were defunded by him too.

[Madelaine McMasters]

6 hours ago, Qie Niangao said:

It's certainly a factor that can be used in analysis, but there are also many other factors that may be relevant to specific policy decisions. I guess I'm arguing that if the associated deaths are tallied in a separate "column", that really needs to be a column in a notional "pivot table" of these factors. For some purposes, living arrangements such as long-term care residence is of interest, sometimes perhaps age, sometimes maybe past exposure to a specific different virus, proximity to chemical plants, etc., which may be (quasi-)independent contributors, each relevant to one policy or another.

Oh yes, there are a lot of variables in this analysis. But it makes little sense to publish, for public discussion, variables around which your public policy is NOT going to pivot. The more I think about this, the more I think a second column for nursing home/prison deaths might be appropriate. We've already seen this anticipated by early release of inmates from prisons, to avoid massive outbreaks. Policy makers were clearing pivoting around something other than the (measured or expected) total COVID death trend line. They were pivoting around a sub-model that predicted a much higher Ro due to crowding in prisons. You'd not invoke the same policy for nursing home residents.

Imagine this in reverse for a small community that's home to a prison or nursing home. You certainly wouldn't schedule the timing of any "stay at home" order around the communities overall trend lines, forcing people to stay home until the prison/nursing home reached herd immunity. You'd separate the high density and low density populations statistically and work out appropriate policy within and between cohorts.

Right now, the aggregate statistics (case, death, recovered) that are bandied about in the news are widely presumed to be driving immediate policy. If that's the case, you'd want to remove from the data set those points that decrease the sensitivity of the problem to your immediate solutions. Since nursing homes, prisons, and other dense housing situations might (I don't know) behave differently in response to stay-at-home, it could be worth sequestering their situation into a different column. I can't make this argument as strongly for other factors you've mentioned, like proximity to chemical plants or other strains of virus. Those are more subtle, less well understood, and less measurable. That said, you know such data is being examined because we're reading news stories telling us about interesting correlations. Imagine the firestorm that would surround local ordinances requiring residents within a certain distance of a chemical plant (which may not be a good proxy for vulnerability) to stay-at-home. With limited ability to identify COVID-19 antibody response, and virtually no ability to measure possibly beneficial exposure to similar viruses, those factors really can't be involved in "stay at home" policy making.

Stay-at-home (the real bone of contention here) is a very blunt and expensive instrument. It's worth making sure that the data driving the decisions is as good as possible. While the political right appears (rightfully in many cases) cold and heartless in their calls to restart the economy, there really is a cost in delaying that. If you think COVID-19 truth is hard to find, try working that problem.

[Qie Niangao]

27 minutes ago, Madelaine McMasters said:

Imagine the firestorm that would surround local ordinances requiring residents within a certain distance of a chemical plant (which may not be a good proxy for vulnerability) to stay-at-home.

Quite so. You may find interesting this bit of the latest Slate Political Gabfest. A question from host Emily Bazelon about recent anti-stay-at-home protests in Michigan leads guest Amanda Ripley to describe an uprising during Milwaukee's 1894 smallpox epidemic in which public health authorities forced hospitalization of largely immigrant residents of high-density housing while wealthier, less densely populated sections were required to self isolate.

I've been fretting lately about what a responsible government would do -- and communicate -- to most efficiently bring "non-essential" workers back into production, get work-from-home employees back to the workplace (assuming that's even a valuable goal), and otherwise gradually restore something like status quo ante.

Once this is all over we'll look back at whatever remains of our economies and, I expect, wish we'd followed South Korea's path, and thank god we didn't follow Sweden's path. But we're at risk of getting the worst of both worlds: investing heavily in the lost productivity of a frozen-in-place economy, only to throw it all away on another wave (or waves) of equally deadly disease -- and can politicians find the will to restore harsh restrictions again when the data warrants? or then punt to pure "herd immunity" and let the corpses pile up Stockholm style?

In any case, one thing I'm pretty sure won't help in this delicate exercise in dynamic optimization with incomplete information: Riling up participants in the economy, already anxious, with incitements to revolt. This is one time where a dictatorship of technocrats would get a better outcome -- even if they miss optimality by a mile.

[Scylla Rhiadra]

It seems to me a given that the more granular and nuanced our models are, the better -- remembering, of course, that our data is desperately incomplete, and that the smaller the sample size, the less trustworthy the data. I work under the (I hope, safe) assumption that science and health specialists are employing models that are granular and nuanced, and that their professional expertise is sufficient to ensure that they are employing and reading them intelligently.

On the larger issue, however, of how data, and its interpretation(s) are presented to both policy makers and the general public, I think that it is fundamentally undemocratic, paternalistic, and potentially dangerous, not to make all of the information available publicly, including raw numbers, where and when it becomes available. To decide that this or that information shouldn't be made public until it has been properly massaged, filtered, interpreted, or whatever, is exactly the kind of approach that nurtures an abiding distrust of "experts."

Which is why we have never needed good communicators and responsible journalists more than we need them now. Obviously, throwing a whole bunch of numbers at people without explanation is not a good answer: it needs to be contextualized and explained, and explained well and clearly. So, for instance, on the Canadian CBC "Coronavirus tracker" web page, it gives all sorts of data about the progress both of the infection and mortality rate across provinces, and at the end includes a caveat under the bolded headline "Daily Numbers Aren't Always Accurate." That's good, but it should be at the top of the page.

In the final analysis, the decisions here have to be made by politicians. And those politicians will in turn be answerable for the choices they make, and the scientific advice that they choose to follow or disregard, to those who elect them. So it is absolutely vital that voters and citizens have access to data and information, and, as importantly, be educated in the ways in which that data and information should be interpreted.

So, there are two prongs to this. One is ensuring that information is made publicly available. And the second is an ongoing and dynamic process of educating the public in how to understand it when it is.

[Luna Bliss]

25 minutes ago, Scylla Rhiadra said:

So, there are two prongs to this. One is ensuring that information is made publicly available. And the second is an ongoing and dynamic process of educating the public in how to understand it when it is.

Privacy laws, which vary from state to state in the US, are preventing some medical knowledge from becoming publicly available. It seems we need to alter privacy laws during extreme circumstances like pandemics where public safety should be more to the forefront.

[Madelaine McMasters]

36 minutes ago, Scylla Rhiadra said:

It seems to me a given that the more granular and nuanced our models are, the better

On the larger issue, however, of how data, and its interpretation(s) are presented to both policy makers and the general public, I think that it is fundamentally undemocratic, paternalistic, and potentially dangerous, not to make all of the information available publicly, including raw numbers, where and when it becomes available. To decide that this or that information shouldn't be made public until it has been properly massaged, filtered, interpreted, or whatever, is exactly the kind of approach that nurtures an abiding distrust of "experts."

I agree with all of this. The data being provided to the public right now, from almost any source I can find... is crap. This is why I think that breaking out the statistics into smaller grains, and explaining the nuance, is a good idea. There's no better way to promote trust in science than to involve people in it.

And let's start with expressing the data as "per capita". I can't understand why we're not doing this, other than because it allows obfuscation. For both governments and mass media, avoiding per capita statistics gives increased latitude for spin. Though Worldometer does show per capita data in their table, it's not the default for sorting. It should be.

[Scylla Rhiadra]

25 minutes ago, Luna Bliss said:

Privacy laws, which vary from state to state in the US, are preventing some medical knowledge from becoming publicly available. It seems we need to alter privacy laws during extreme circumstances like pandemics where public safety should be more to the forefront.

It's hard to imagine that the kind of anonymized statistical data we're talking about would fall under such a rubric, but if it is, then . . . yes.

[Scylla Rhiadra]

9 minutes ago, Madelaine McMasters said:

I agree with all of this. The data being provided to the public right now, from almost any source I can find... is crap. This is why I think that breaking out the statistics into smaller grains, and explaining the nuance, is a good idea. There's no better way to promote trust in science than to involve people in it.

And let's start with expressing the data as "per capita". I can't understand why we're not doing this, other than because it allows obfuscation. For both governments and mass media, avoiding per capita statistics gives increased latitude for spin. Though Worldometer does show per capita data in their table, it's not the default for sorting. It should be.

And I agree with all of this. (Aren't we reasonable???) Especially this: "There's no better way to promote trust in science than to involve people in it."

I've noticed some evolution in the way that the data is being reported, from at least a few sources. One Canadian source, The Toronto Star (one of Canada's two highest-circulation papers) was annoying the hell of out of me last week particularly because it tended to report only raw numbers, couched in terms like "Ontario today saw the largest single-day rise in number of infections and deaths" . . . all the while ignoring that the rate of growth was actually going down. They seem to have got their act together somewhat this week, and while they are still reporting raw data, they are now noting how the numbers relate to changing rates of infection and mortality. So maybe at least some of the more responsible media sources are becoming more educated?

The per capita data on Worldometer is very good and useful: It's actually my go-to site for data for that reason, as well as the fact that it allows you to see the previous day's data (as misleading as a day-to-day comparison can obviously be). I'm actually not sure how concretely useful a country-by-country comparison likely is, though: the conditions and contexts of Italy and the US are, for instance, hugely different. I sometimes have the sense that we are running a sort of Covid-19 Olympics, with nations competing against each other for the most successful approach to the virus. At the same time, of course, I understand that the experience of places like Japan, Korea, China, Italy, Germany, etc. are really useful resources for calculating strategic responses. I just think think it's not very helpful to look for bragging rights.

[Luna Bliss]

9 minutes ago, Scylla Rhiadra said:

35 minutes ago, Luna Bliss said:

Privacy laws, which vary from state to state in the US, are preventing some medical knowledge from becoming publicly available. It seems we need to alter privacy laws during extreme circumstances like pandemics where public safety should be more to the forefront.

It's hard to imagine that the kind of anonymized statistical data we're talking about would fall under such a rubric, but if it is, then . . . yes.

Here's some info about it:

https://www.usatoday.com/story/news/2020/04/13/coronavirus-nursing-homes-2-300-facilities-report-positive-cases/2978662001/

An employee in a nursing home was not even informed that Covid-19 was present.  I've read stories of carers abandoning some nursing homes. It's a real mess.

 

It appears some states misinterpret the medical privacy laws contained in HIPAA:

https://www.poynter.org/reporting-editing/2020/hipaa-and-coronavirus/

[Madelaine McMasters]

13 minutes ago, Scylla Rhiadra said:

I've noticed some evolution in the way that the data is being reported, from at least a few sources. One Canadian source, The Toronto Star (one of Canada's two highest-circulation papers) was annoying the hell of out of me last week particularly because it tended to report only raw numbers, couched in terms like "Ontario today saw the largest single-day rise in number of infections and deaths" . . . all the while ignoring that the rate of growth was actually going down. They seem to have got their act together somewhat this week, and while they are still reporting raw data, they are now noting how the numbers relate to changing rates of infection and mortality. So maybe at least some of the more responsible media sources are becoming more educated?

This is the "first derivative" problem we discussed recently. Looking at the total case/death counts isn't terribly helpful if you're trying to decelerate. If you're driving down a steep hill and worry your brakes will burn out, you look at your speedometer, not your odometer. This is apparently a difficult concept for people to grasp.

18 minutes ago, Scylla Rhiadra said:

The per capita data on Worldometer is very good and useful: It's actually my go-to site for data for that reason, as well as the fact that it allows you to see the previous day's data (as misleading as a day-to-day comparison can obviously be). I'm actually not sure how concretely useful a country-by-country comparison likely is, though: the conditions and contexts of Italy and the US are, for instance, hugely different.

I'm also not sure how useful per-capita data is, but I'm certain it's more useful than total counts.

22 minutes ago, Scylla Rhiadra said:

I sometimes have the sense that we are running a sort of Covid-19 Olympics, with nations competing against each other for the most successful approach to the virus. At the same time, of course, I understand that the experience of places like Japan, Korea, China, Italy, Germany, etc. are really useful resources for calculating strategic responses. I just think think it's not very helpful to look for bragging rights.

We really are running an Olympics, aren't we? While the "scores" recorded on Worldometer are incomplete and potentially misleading, there will eventually be prizes awarded to those countries that performed "best". Twenty years from now, if there was a shift in the balance of power across the globe resulting from COVID-19, I'm sure winners will be declared and disputed. Now that you've drawn this sports analogy, I'm wondering whether it's a competition against the clock or tape measure, or a competition between teams. I see elements of both.

[CaerolleClaudel]

So, this is the opposite of venting, I guess, but maybe that is ok? 🙂

I am reminded yet again today of how much I crush on Jacinda Ardern. She has done so many awesome things since the took over, and is putting on a master class with her COVID messaging and leadership, pretty much the opposite of the approach taken by some orange obese dotards 'leading' some countries. I realize NZ is a far smaller and I am guessing more homogeneous place, but she is just killing it, IMO.

Plus, wow, she is just tooooo pretty...smart, capable, and pretty! 😍

[Madelaine McMasters]

2 hours ago, Qie Niangao said:

Once this is all over we'll look back at whatever remains of our economies and, I expect, wish we'd followed South Korea's path, and thank god we didn't follow Sweden's path.

I've been discussing Sweden with a RL colleague who's spouse lives there. All three of us are numbers junkies. They'd been arguing that Sweden was waltzing into a disaster. At the start of our discussion, I took (as I do) the contradictory stance, citing Worldometer data as the source of my doubt. It's been about 10 days since then and the data is working against my theory. The data is also peculiar...

That chart is a classic example of measuring your tools as much as measuring the actual thing, and doesn't give me much confidence in Sweden's data.

Edited April 19, 2020 by Madelaine McMasters

[Luna Bliss]

13 minutes ago, Madelaine McMasters said:

the data is working against my theory

Have you tried remote-viewing or tuning into the Akashic Records to get better data?