The curve is flattened enough

The Quarantine Zone
81

Looking for resource. (Think of this as an old school classified ad).

I would love to see a list of what relaxation measures different states and countries have taken when, and for each what has happened to their case rates, percent positive rates, hospitalization and then ICU rates over the following two to four weeks, as it’s been that long. Which ones reimposed quickly and which ones stayed without major new flares? Yes careful interpretation for specifics will be needed but maybe some patterns can emerge.

Anyone have a resource?

82

Appears to have some of what you’re looking for. I haven’t spent much time on the site and make no claims on reliability of data, but it appears OK.

83

i wasn’t suggesting that the specific measures that worked to control HIV would be applicable to SARS-CoV-2; rather that there are options besides either waiting for a working vaccine (which, despite recent optimism, may be years in the offing) or just giving in and letting the infection take course without remit. Even if immunity fades quickly and reinfection can occur, it is clear that for a large majority of people this is a survivable illness without treatment or hospitalization, and with better testing and surveillance we may be able to control outbreaks such that they do not reach epidemic proportions. This would require modification to how we conduct business and social activities, but as has been discussed previously, we should be doing so anyway because the next pandemic may be far more virulent than this one, and it is out there in our globally connected future.

I don’t have a single unified resource but there is the Johns Hopkins Coronavirus Resource Center (which seems to be the most up-to-date resource on COVID-19 epidemiological data), the Wall Street Journal “A Guide to State Coronavirus Reopenings and Lockdowns” (which has an accessible summary of state lockdown and relaxation orders), and The Council of State Governments “COVID-19 Resources for State Leaders” (which has an extensive listing of executive orders, state reopening plans, and links to the major predictive modeling efforts).

I also find it instructive to listen to the This Week in Virology podcast (which is obviously all about SARS-CoV-2 recently) because although while they are more focused on pathogenesis than epidemiology they often cover some very salient issues, such as the fact that a positive RT-PCR test only indicates the presence of viral RNA fragments (which can remain residual in a previously infected person for weeks or months) but not necessarily active virus, so all of these supposed positive retests are not definitive unless researchers run a plaque assay. Be warned, TWiV is a long form podcast (often in excess of two hours) with a lot of technical jabber, but at least on Apple there is some timestamping that lets you jump to sections of interest.

Stranger

84

This is a pretty good visualization of what is happening on a state-by-state level, if the data is accurate. It’s interesting scrolling thru the states - the graphs show when Shelter-In-Place orders were issued (if they were issued), when reopening, as well as recent 14-day trend lines for cases and deaths.

You can see some states have flattened the curve pretty well, others are still climbing, but even if cases are still climbing (as expected) deaths are generally trending down.

85

I followed Trom’s link, and I can’t help but feel sad about Virginia.

86

70% is an assessment based on a presumption predicated on presumptions. With some extrapolation from similar viruses.
Less contagious viruses can mean here immunity at 10-15%. More contagious ones reach immunity at 90% and some are so contagious (chicken pox, measles) that herd immunity is functionally impossible to reach.
Spanish flu of 1918 saw herd immunity kick in at about 50%.

And it’s not like that you go from fucked to fine as we increase from 69%- 70%. As the number of immune people grow, it has a harder time spreading.

The current numbers presumes the population has no immunity or resistance, which is unlikely to be the case, one paper suggest that exposure to certain strains of cold causing coronaviruses provides some immunity and resistance to COVID-19.

https://www.cell.com/cell/fulltext/S0092-8674(20)30610-3

Which might go someways to explain the vastly different Death and morbidity rates amongst similarly positioned countries.

87

I thought the flattened curve really meant effectively, that now, they have room for you at the hospital.

88

The goal is also to flatten to buy time for there to be effective therapeutics. Someone who gets the virus six months from now likely has more options than someone who gets it today.

89

Not necessarily. It literally means the curve has been suppressed to some extent.

Hopefully it has been suppressed to the extent that our hospitals aren’t overwhelmed but that’s something that’s very location dependent.

There tends a lot less slack in rural areas than urban because of the relative lack of hospital beds per capita, so even within the same state the same flattened state-level curve may be good for some people but not enough for others. Hopefully somebody is looking at those numbers as well.

90

“No way to prevent this says only country where this happens regularly”

Before you decide that there is no way to avoid the virus blasting through the population, maybe take a look at the curves of other nations. See what they have achieved without vaccines or antivirals, and in many cases with very weak healthcare systems and poor testing and tracking capacity.

Just with lockdowns and social distancing.

(I don’t know what criteria that site uses, or why Bhutan is “winning” while south Korea is not. But it is a good collection of graphs, and they are the information I wanted to convey.)

91

America is ‘different’. Nothing that works in any other part of the world can possibly work here for … reasons.

CMC fnord!

92

Links appreciated but unfortunately none seem to my first looks to be what I am hoping for, albeit they are decent resources to contribute to creating it if I have the energy. Which I clearly don’t have today. But if I want it I may need to at some point I guess.

All these states and countries are relaxing measures in different ways. I’m looking for a one site resource that compares and contrasts the different approaches and their results on meaningful metrics at meaningful intervals to the changes. Not just “new cases”.

It seems like something important enough for someone else to have already done.

93

That sounds like a good idea on the face of it but in order to actually have “meaningful metrics” or what I would term figures of merit you would have to have both a way of characterizing relaxation measures in a discrete fashion, and have some way to calibrate the relative efficacy which includes things like variance for socioeconomic factors, social and commercial compliance with isolation orders, access to healthcare, the effect of political leaders reinforcing or undermining directives, et cetera, lest all of this ‘noise’ overwhelm the actual trends of effectiveness that you are looking for.

I don’t think this is impossible but it would take a lot of skilled understanding of these factors to really tease out actual effects from common measures from the individual variance between states (similar to what epidemiologists and social researchers do for studies on nutrition, education, et cetera) and notwithstanding that there are a lot of problems with testing and data reporting which means that in some places there are certainly underreporting both actual infections and deaths (not attributed to COVID-19 because the occur at home or the patient has a significant pre-existing condition). At this point, I think the best that can be done is lump together states by region and then make comparisons between individual states. A single “lockdown protest” by a few thousand people that results in a flurry of new infections because nobody was isolating (and which shows up two or three weeks afterward) and people returned to their communities and spread contagion from there is enough to significantly skew results, and that is something that no model is going to predict or account for. What would really be useful from a data analysis standpoint would be individual tracking data so you could have a big bag of data in the aggregate that would directly show how much compliance was achieved and where people went, but that has obvious severe privacy implications even if the data were stripped of identifying metadata.

Stranger

94

Stranger - no question the interpretations would be tentative, because in fact it is not only America that is different, every place has its quirks. But still useful to look for patterns that emerge, and minimally some experience if certain items that some claim are absolute pan out.

Just starting a wee little bit I began with Austria since they started sooner. Total deaths plateaued at under 71/million. Probably would be best to scan psychonaut’s posts, but went with Googling:

May 1, almost 3 weeks ago, opened up to letting people outside without masks keeping 3 feet away, facemasks in enclosed public spaces, gatherings up to 10 people, funerals up to 30, shops open, restaurants 5/15 c 3 ft between tables of 4 and servers with mask, churches with distancing and masks, schools were scheduled to open 5/18 in split shifts but I see no reports of whether or not they did.

A contact tracing app is available but not used too much. Currently about the same number of tests/million run as the United States. But therefore a much lower percent positives.

Definitely lots of chomping at the compliance bit even with loosening. Yes protests and a narrow majority saying they would be against new restrictions in the case of another wave. Still most wear masks often even where not legally required to do so, even though most there doubt they do much.

A few blips as reported by **psychonaut[/b, the disease is not gone, but the moving average of daily new cases has not jumped up and the new deaths rolling average is staying below 1. Long enough for the initial opening to be having impacts, not long enough to see about restaurants, churches, and schools.
What I am hoping someone else has done is that for all the early opening countries and any state that has opened for at least 2 weeks. But just doing Austria took me enough time for now!

I know Italy has kept rates dropping with some opening since mid April and much more since 5/3 including parks and manufacturing. And Denmark has had schools open over a month. But for each we’d need to know the whole package, where deaths were when they opened, and what key metrics have done since.

Which went too far too fast for their specifics and had to pull back? Korea, but they opened bars! China has had to reimpose some restrictions in some locations. Lebanon after some of their citizens returned from abroad. Just today Saudi Arabia. Iran one province.

Put them all together and you have a bit more to base a guess of how much flattening allows you how much opening with how much bounce back.

95

Not necessarily. You want the curve of hospitalizations to be both flat and significantly below capacity such that both you aren’t running up against resource limits (e.g. you aren’t going to run out of PPE, or use up needed pharmaceuticals, or burn through skilled medical personnel) and you have reserve capacity to deal with a surge. Ideally you are managing infections so the curve is always below the threshold and slightly downward, anticipating that some change will force the slope to positive but with enough lag that you can detect it.

Absolutely. We need to buy time to figure out what works, and also to replenish supplies, train new medical personnel, and generally give people on the front lines a break because while they are being venerated as “heroes”, they are people with their own families, fears and limits of endurance and stress tolerance. If researchers can come up with some therapeutics that keep people from having to go weeks on oxygen support or ventilation, the number of patients that can be treated is much greater and especially in rural areas with only moderately equipped regional hospitals that were already stretched financially and resource-wise even before this pandemic.

You can certainly use them to refine general guidelines; I’m just dubious that you can actually make predictive trends that are significantly better than the more generic models that are currently in use. If I were going go to into data scientist mode, I’d pick certain measures like restricting large gatherings, closing certain types of business, closing schools, et cetera, ascribe a kernel density estimate (KDE) to each of them (and maybe a scale factor based upon the relative restrictiveness) and then use them as weighting values on an assumed nominal distribution of R[SUB]0[/SUB] and then look at how the resulting replication number changes based upon the data and go back and modify the KDEs accordingly. By looking at the posterior values you can figure out what the relative significance of different measures are and what the most likely value actually is for the modified R[SUB]0[/SUB], and then try to adjust the measures to keep the replication number below unity but high enough that you are getting progressive exposure and inoculation such that as time goes on you can open up more and more but still keep contagion to sub-epidemic levels.

However, the latency of contagion, the errors and delays in data reporting, variation in compliance, and just the speed at which this virus can spread are all going to make it challenging to really dial in any response. Erring toward “just enough to keep the economy moving” is probably as good as can be managed, and likely not enough to satisfy the desire of the public to get back to a more normal society, so we are almost certainly going to be seeing epidemic-scale outbreaks even in states that are applying more cautious opening plans.

Stranger

96

Given how similar all those curves look, maybe it’s not ‘just lockdowns and social distancing’.

97

Did you actually go to the link? Do Sweden and Taiwan look similar?

98

Not being a data scientist I’d be not so fancy! :slight_smile:

I’d just pick a hypothesis and see if the data supports it or not.

To pick one at not so random, the hypothesis that keeping schools closed is a critical part of keeping rates low. If several re-openings have included re-opening schools and daycares without seeing any resurgence in new case numbers, increased percent positive cases, hospitalization rates, ICU utilization rates, or death rates, at appropriate time lags, and the places that have had flares and had to reimpose restrictions did not mostly share that they re-opened schools, then the hypothesis begins to become falsified. A big deal thing. Of course you need to look at each case to see what else might be correlated with it but it still informs.

Alternatively pick the hypothesis that keeping bars closed matters: South Korea’s experience supports that strongly.

As powerful as your approach? Maybe not. But still useful.

99

Skepticism is good, but the central point is still valid - Nearly a month after Georgia began reopening, there is no evidence of the widely anticipated ‘spike’ in new cases.

100

Yes, I did. That’s how I was able to observe that they all looked similar.