Morgan3820
New member
Makes me want to move to Texas.
https://amp.cnn.com/cnn/2020/03/23/...texas-social-distancing-guidelines/index.html
https://amp.cnn.com/cnn/2020/03/23/...texas-social-distancing-guidelines/index.html
Finally a voice of reason
- Thread starter Morgan3820
- Start date
Yes, perhaps easiest to just plot the log of the number of total cases. If that looks like a straight line, then it is exponential growth at a constant rate.Palmpilot said:
As of the day before yesterday (when I last did this for the US data), it looked like a fairly constant exponential growth with a 2.5 day doubling time. There was a slight hint of a decrease in the slope, but nothing that appeared outside the range of random variation.
There is fairly good accumulating evidence that hydroxychloroquine shortens the duration of viral shedding to between 4-6 days.Tantalum said:
Whether such people are then immune is not presently measured, but that would be the normal result.
If Covid-19 becomes a trip to the doctor and some illness for a week or two (like many viral illnesses), with a need to take medicine but recovery following, it is a much less dangerous thing.
No, I am not aware of a trial in NY (which is not to say there aren't any, I just haven't seen them published). Always happy to provide good citations when possible.Chip Sylverne said:
Here is a recent one out of China - http://subject.med.wanfangdata.com.cn/UpLoad/Files/202003/43f8625d4dc74e42bbcf24795de1c77c.pdf
Here is an older one out of China for the related drug Chloroquine phosphate - https://www.jstage.jst.go.jp/article/bst/14/1/14_2020.01047/_article
Here is a study out of France which had fair controls but a small sample size - https://drive.google.com/file/d/186Bel9RqfsmEx55FDum4xY_IlWSHnGbj/view
And here is that Lancet article - https://www.thelancet.com/journals/laninf/article/PIIS1473-3099(20)30132-8/fulltext, which is really not a clinical trial but suggestions for possibly useful drugs.
To somewhat keep up on these, I search each day on scholar.google.com for 'covid-19 treatment' and restrict results to 2020. Most of interesting results in the first 3 pages. I know that a lot of physicians have apparently been prescribing hydroxyquinolone, apparently so much so as to cause a temporary shortage of it at pharmacies.
If people are interested in following this more, I post an update every day or two on my MeWe page which includes any trends in total cases in the US as well as possible treatments being developed. You can add me as a contact if already on MeWe or join and add me. mewe.com/i/petersteinmetz
Some confusion these days may stem from another common current usage. That is when people say one number is exponentially larger than another, with no implication of a change in time or with respect to an independent variable.jimhorner said:
That one is a pet peeve of mine and just wrong in my opinion.
Yes, that is what is meant by exponential growth. The rate of growth appears to have changed in a statistically significant way since March 1st around March 25th.Sluggo63 said:
I would use some caution in ascribing causation here at least at the national level (where I have looked at the numbers). I can think off hand of 4 reasons why the exponential growth rate of the number of total cases might decrease (which it does appear to have done in a statistically significant way around March 25.Tantalum said:
1. The measures overall taken across the US, such as shelter in place, lockdowns, and just the advice for social distancing, may have worked.
2. It is getting warmer and perhaps the virus is sensitive to temperature as many other seasonal flus are.
3. This could be some weird function of testing availability. Perhaps an exponential growth in cases exhausted extant testing supplies and now a more linear increase in the number of available test kits is dominating the overall effect.
4. It may be that number of asymptomatic cases really dominates the symptomatic ones, by some estimates a factor of 5, and as per the Oxford modeling study a large fraction of the population has already been exposed and we are simply seeing a slowdown in the rate of the number of cases that get noticed and tested.
Overall likely good news, but I think it will take some time before people are able to parse out what may have helped here in the US.
I missed this earlier that the Apple News item is the item based on the Kinsa temperature readings. This is very weak evidence because of large potential confounds. It is entirely possible that if you tell people there is a terrible pandemic and that we having to go to lockdowns, that one of the things people do is take their temperature because they are afraid they have this terrible pandemic. And if a large number of them are not in fact infected, the average temperature measured will drop.
Could be, as in explanation #3 in my post. But it does seem odd. Do you know of any datasets describing the availability of tests as a function of time? That would be quite interesting.weilke said:
Yes, most seasonal flus die down in the summer. There are several hypotheses for why that is the case, amongst them warmer temperatures, though there are others, such as people spending less time indoors and kids being out of school.Tantalum said:
OTOH, it is unclear that SARS and MERS followed that pattern, so since those are more closely related to Covid-19, Covid-19 may not follow a seasonal pattern.
I suspect it will take quite a bit of research after the fact to really pin down which factors were causative and by how much. At lot we don't know about this one at this point.
Would you care to speculate on what process causes that approach to the limit of cases in Italy? That intrigues me. That is a log plot and it looks like the log of the difference between the number of cases and the limit is decreasing exponentially. If it was a log decay to the limit of the number of cases, that would seem to indicate an exhaustion of the available non-infected people. But in this case, it is an approach on the plot of the logarithm.weilke said:
It is a good temporal observation, but as in my post above, I would be a bit cautious about assuming causation. There are other potential explanations and the lack of knowing generally how many people have been exposed and when is really hurting the ability to make good inferences at this point. In the UK for example, a modeling study at Oxford suggested that in fact over 50% of the population has already been exposed and what we are seeing in their numbers is some type of function of testing bias. The estimates of the number of asymptomatic cases in the literature have varied from 20-86% and clearly that has a big impact on interpretation here.chemgeek said:
Whether governors should institute lockdown policies based on the current data is actually a separate political question of how certain one has to be that a person infected with Covid-19 is a danger to others, but that is likely to lead to thread lock here. (I suspect the moderators are just tolerating Covid-19 since it is of great interest to the community, though has little to do with aviation.)
It is a better more certain measure for sure. But my point is one can’t assume causation from either measure and a temporal correlation. A lot of variables in play here and very poor measurements of fundamentals like rate of showing symptoms or whether there is even excess mortality due to Covid-19 (some data from Europe that Covid-19 displaces other deaths but has not contributed to an increase in overall deaths).chemgeek said:
Isolation measures contributed to decreased Covid-19 deaths seems like a good first hypothesis, I agree, but I suspect it will be a study for some time to try and figure out what the causal factors were and how effective each was.
Was thinking about it more and this is in general an important observation. If you lock everyone in their house for a month, the propagation of this will stop. The problem is, everyone will have starved to death. Similar to how if you ban all GA flying, there will be no more deaths due to GA accidents.Tantalum said:
There is a tradeoff involved at less absurd levels of intervention, and a fundamental question, though a political one so I will stop after noting it, is what is the appropriate tradeoff between other deaths and other costs and Covid-19 deaths and illnesses? Not a question we have much good data to answer presently, but there is a tradeoff there in reality whether we choose to acknowledge it or not.
Another item I will share from something I have been tracking -- which is total US confirmed cases and deaths. It appears from some basic statistical testing that the rate of exponential growth started decreasing in a significant manner about March 25. While deaths in the US has accelerated a bit today, 14 days from then will be April 8th. Let's hope the deaths decelerate then for the US as a whole also.
You are correct, the normal seasonal flu is not a coronavirus. Nonetheless, many viral infections follow a seasonal pattern, with infections lower in the summer. As you note, some of these are colds, which are coronaviruses. Pretty active debate amongst experts right now if COVID-19 will follow a seasonal pattern. We can hope but remain vigilant.Cap'n Jack said:
Sure, always happy to provide pointers to what I am looking at it. I will note that this is fairly new and could turn out to be weird variation.tspear said:
This has been noted in the UK (https://www.ft.com/content/f3796baf-e4f0-4862-8887-d09c7f706553).
And then more recently the latest European figures show excess mortality for Italy but not Europe overall where mortality is actually lower in the last reported week and that for flus overall is not so high as in prior years. (https://www.euromomo.eu/index.html).
Other commenters have noted that this seems quite surprising if it is true and continues. One possible explanation is that the measures which have been taken have reduced overall mortality, offsetting any increase in Covid-19 deaths. Another is that Covid-19 deaths are largely just displacing other causes of death in susceptible individuals. Or perhaps it is a combination of a number of such factors.
I am very curious to see what similar numbers in the US are like. Very preliminary and sort of puzzling.
Strange about FT because I can access that article without a subscription.tspear said:
I don’t quite understand what you mean by “you need a pandemic...” Perhaps you could expand?
The big problem with the modeling right now is that our knowledge of the underlying parameters is so limited presently.
That is why I find this data so intriguing. Why isn’t there a big spike in excess deaths? If there is an increase in Covid-19 deaths but no increase in overall mortality, what happened?
Or do you mean that the number of Covid-19 deaths is so small as to be unnoticeable in the total number of deaths? If this latter, I guess that brings me to the next point.
As I noted before, how people want to trade off Covid-19 deaths against other deaths or resources is a political question. There is certainly a non-zero cost associated with interventions to reduce Covid-19 deaths, either in terms of money or deaths due to other causes — it is not zero. And with all such costs, different people will have different preferences. But that balance is a political question that I would suggest we avoid here to avoid thread lock.
Well, as noted above, the number of deaths tends to be a more reliable number, though assignment of cause(s) is clearly a softer call.weilke said:
But in terms of the numbers, if Covid-19 deaths are up substantially and overall mortality is down, what happened? Or say overall mortality remains the same, what happened?
Presumably either some deaths were just relabeled essentially as Covid-19 or in fact other deaths are down to offset the Covid-19 deaths.
I don’t see the latter as particularly improbable. If you force people to stay home, they are probably not engaging in a number of activities of daily living which otherwise kill them - like driving cars, getting drunk in bars and fighting, etc.
Of course this may all be a weird anomaly in the data at this point - maybe some underreporting? We need to see how it develops.
I guess I don’t understand what you think is specious -perhaps expand?
I am puzzled by citing that article in particular, given the last statement here. That article’s abstract states near the end, referring to another paper “The authors demonstrate an impressive statistical association between vapor pressure, influenza transmission, and virus survival.” None of these things seem to relate directly to social distancing and most of the article is discussing other extrinsic factors. The summarized items about incidence during the school year are a relatively minor part of the article, almost an aside. The article is also from 2009, which doesn’t mean it is incorrect, but suggests there may be newer research. I suspect there may be newer articles, or better yet, reviews, which deal with the relative causation of seasonal cycles in influenza.wrbix said:
In any case, returning to the subject of whether Covid-19 may be seasonal, I have seen academic articles arguing for and against that proposition. At lot to be learned about this.
