PlagueWatch 2020-23 - Coronavirus

[Acymetric]

Yes, they spread the blame around...what was concerning was the assertion from the young woman with cancer that Duke student health is widely viewed as being "awful." I wonder if that's true, I have no way of knowing.

My experiences aren't at Duke (or a school that had a hospital/medical program), but I think the default assumption for student health services is that they aren't very good. One might hope that a place like Duke with their medical system would be better, but I wouldn't necessarily assume that.

[DukieInKansas]

It was true when I was there. And my experience with it only confirmed the notion. I had an icky experience there as a senior too. I'll defend Duke about a lot of things, but their student health center, not so much. That said, it's a problem with student health centers in general. I get the feeling that they all believe every student is coming in either because they are drunk or trying to get out of exams. And let's face it, they probably see a lot that kind of thing. Although they ought to be better at recognizing mono. They probably see a lot of that too.

To be completely fair though, they did a decent job taking care of me when I broke 3 toes my junior year.

I knew a student that didn't want to go because the typical response to colds was sudafed and robitussin. I assured her that, if they were told that you got the same thing every year and your family doctor prescribed x, then they would give x. She was pleasantly surprised when they did just that.
I only went once, for something that happened periodically, and they gave me what had worked for me in the past - but suggested I should see a counselor (stress exacerbated allergic reaction - that happened maybe once a year.) I'm with BD - probably saw a lot and listened if the student talked to them calmly about their needs.

Is there an interesting story to go with the 3 broken toes?

[freshmanjs]

I am in favor of using information to guide decisions and revising past decisions as new information becomes available. I doubt you’ll get much argument with that position, for what it’s worth.

People won't argue with the statement as long as it's abstract. In practice, people get very uncomfortable / angry if data is presented that could be threatening to what they think are the necessary or correct behaviors or decisions.

But, of course, decisions have to be made.

Based on your research, if you were making a decision today, and that decision require you to assume a hospitalization rate for Covid-19 compared to flu, would you assume that rate is higher than, equal to or below?

Per case hospitalization rate? I would assume it's similar.

Flu hospitalizations per case are somewhere between 1-2% based on the CDC data I've seen. So, if you have 450k Covid-19 hospitalizations to-date (roughly), you would have to believe we've had 22-45M Covid-19 cases for the rates to be similar. I don't think that's out of the question. We've seen estimates that only 10% of Covid-19 cases have been identified.

The things we don't know and probably matter more than per case hospitalization rate are:
- How many cases would we have under various policy scenarios? In the extreme, if you do nothing, how many people will get Covid-19 per year? It is not 100%. How many is it? We don't know, although the Florida and Texas situations are giving us more insight unfortunately
- How is the hospitalization rate changing as the disease mutates, exposures are more limited, weather changes, etc? This is not clear yet, but also important

For now, I think cautious reopening with widespread mask use is the right approach for most places in the US. If only we could get our act together...

[cato]

Per case hospitalization rate? I would assume it's similar.

No. Absolute numbers. The kind one needs to consider if making decisions.

[freshmanjs]

No. Absolute numbers. The kind one needs to consider if making decisions.

Hmm...I make data driven decisions for a living and typically use things like per case numbers and relative numbers much more often than isolated absolute numbers.

To answer your question re: absolute numbers for different time periods:

- October 2019 through today, I think flu is higher but it's close.
- Weekly rate at peak of flu season compared to peak of Covid-19 in April, Covid was higher.
- Current weekly run-rate for Covid vs. peak flu season weekly rate, Covid is probably higher
- For the next 12 months August 2020 - July 2021: it depends on policy and behaviors. My best guess is flu will be higher.

[MChambers]

Here's an article written by a physician who had a patient contract Covid-19 twice. It was worse the second time.

Double yikes.

https://www.vox.com/2020/7/12/213216...-herd-immunity

[Matches]

My ex-wife’s father (my kids’ grandfather) has tested positive. He is asymptomatic and hopefully will remain that way. The kids have not been around him recently though their aunt has. Fingers crossed.

[CDu]

As far as what is causing deaths to be lower right now despite the huge increase in cases? I think it is due to a few reasons:

1. We are testing MUCH more now than we were in March and April. That means we are catching more of the asymptomatic/lightly symptomatic cases. In the early stages, individuals with minor symptoms we’re often told to stay home in order to save the limited number of tests for more severe cases. Also, in the earliest stages, we simply didn’t have the means. I would suspect that our case counts in March and April were probably 10-fold too low.
2. Relatedly, by testing more, we are probably catching cases earlier in the disease pathway. The earlier a case is caught, the longer it will take from “observed onset” to resolution. If you don’t catch cases until symptoms manifest as severe enough to require hospitalization, time to death will be shorter than if you identify cases still in the outpatient setting.
3. The death rate really is going down. This is, of course, a conceptually straightforward idea. In the early stages, before we realized what was happening, the virus hit hard everywhere in the populations it reached. And “everywhere” included the less healthy, high-risk populations. So deaths spiked. Then, we implemented lockdowns, and those in the high-risk groups started taking extra precautions. So the average age of infections has decreased, which reduces the case mortality rate. I would venture that the comorbidity profiles are better among the infected now too, which will also help the case mortality rate.
4. Perhaps a small part is better treatment. Remdesivir is being used in severe cases. Dexamethasone too. These have been shown to reduce mortality. They aren’t in widespread use, but every little bit helps.
5. System overload has - to this point - been avoided since the first spike in deaths. When hospital capacity is exceeded, outcomes get worse. After the lockdown, hospital burden decreased. While things have spiked again, it is happening over a larger swath of the population this time. So we haven’t yet hit system overload again. Though it sounds like we could be closing in on that in some places.

If I can find time, I will do an overlay of the timelines of observed cases, a hypothetical “actual” cases, and deaths to illustrate points 1 and 2. I tend to think 3 and 5 are the biggest drivers. Hopefully 3 is the biggest, but I can’t yet rule out 1, 2, and 5 as being significant factors..

[rsvman]

I don't think it is fair to use hospitalization alone as a marker for severity. It seems to me that a higher percentage of covid patients who end up in the hospital are generally more severely affected, based on ICU numbers. Hospltalization, per se, is inherently a more arbitrary measure of severity than ICU admission or intubation and mechanical ventilation.
For example, many flu patients end up in the hospital because they have myositis, and thus need aggressive hydration and monitoring to keep them from developing kidney failure. Most of these people, although requiring hospitalization, do not require admission to intensive care, and can be discharged within a few days.

I don't remember a flu season ever in which the ICUs became filled up and other parts of the hospital had to be utilized as overflow intensive care, or where extra ventilators had to be acquired, or where the Navy had to send a ship up to the Hudson river to take care of overflow patients. If it has happened, I have not been informed of it.

So, at a minimum, I'd have to say that the rapidity with which this virus spread, and the numbers of extremely severe cases is beyind what we have seen with infuenza in the past 50 years, at least, even if the total hospitalization numbers are similar at this point in time.

[cato]

Hmm...I make data driven decisions for a living and typically use things like per case numbers and relative numbers much more often than isolated absolute numbers.

Curious as to when you use absolute numbers instead of relative? I am thinking absolute numbers are important when analyzing bottlenecks/failure points, like the ability to treat patients requiring hospitalization.

The thing about dealing with an unprecedented disease is that the ways we typically make decisions may not be the best ways to make real time decisions as the pandemic is unfolding.

To answer your question re: absolute numbers for different time periods:

- October 2019 through today, I think flu is higher but it's close.

Based on this, I would assume that hospitalizations from Covid will be higher than flu for at least the rest of this quarter. Why? If flu and Covid were roughly the same from October 2019, before Covid was even on the scene, through today, when it is just solidifying it’s hold, then this will likely trend worst for the immediate future.

- Weekly rate at peak of flu season compared to peak of Covid-19 in April, Covid was higher.
- Current weekly run-rate for Covid vs. peak flu season weekly rate, Covid is probably higher
- For the next 12 months August 2020 - July 2021: it depends on policy and behaviors. My best guess is flu will be higher.

I’m curious how you reach that best guess? We’ve learned how to protect the most vulnerable from Covid and at some point in the next 12 months we will have a vaccine that is at least as effective as the flu vaccine?

[CDu]

I don't think it is fair to use hospitalization alone as a marker for severity. It seems to me that a higher percentage of covid patients who end up in the hospital are generally more severely affected, based on ICU numbers. Hospltalization, per se, is inherently a more arbitrary measure of severity than ICU admission or intubation and mechanical ventilation.
For example, many flu patients end up in the hospital because they have myositis, and thus need aggressive hydration and monitoring to keep them from developing kidney failure. Most of these people, although requiring hospitalization, do not require admission to intensive care, and can be discharged within a few days.

I don't remember a flu season ever in which the ICUs became filled up and other parts of the hospital had to be utilized as overflow intensive care, or where extra ventilators had to be acquired, or where the Navy had to send a ship up to the Hudson river to take care of overflow patients. If it has happened, I have not been informed of it.

So, at a minimum, I'd have to say that the rapidity with which this virus spread, and the numbers of extremely severe cases is beyind what we have seen with infuenza in the past 50 years, at least, even if the total hospitalization numbers are similar at this point in time.

The ICU point is a good point: ICU cases per hospitalization is probably a better measure of severity. Per that same MMWR study, a bit over 16% of COVID hospitalizations reached the ICU. Comparatively, only about 6% of flu hospitalizations ended up in the ICU per this study: https://www.healio.com/news/infectio...-icu-admission

Your last paragraph is important too: the simple reality that systems were overwhelmed in ways that haven't happened with the flu suggests that COVID is worse than the flu.

[Tappan Zee Devil]

Originally Posted by Phredd3
"Barbaric is a strong word, but it may not be a great decision. The study you cite looks entirely at the risk of death, and as folks have repeatedly pointed out in this thread, death is not the only bad outcome to COVID-19. Significant hospital stays and potentially long-term effects to the disease are not especially uncommon. Or, to quote another article posted up-thread:


I don't know about your school, but my kids' high school is absolutely loaded with people in the 50s age range, including many of the very best teachers in the district. Personally, I'm not happy to subject them to that. Also, I'm not terribly willing to run the experiment with my high-schoolers, given that, according to retrospective studies in France, China, and Iceland, the secondary infection rate for households of high-school-age children is 10%-15%. I'm really not all that happy to subject myself to that substantial of a chance to contract this kind of infection.

Things are quite different at the elementary level, where transmission among children 11 and under seems to be quite uncommon. One size does not fit all, even when it comes to schools."

Yes but the question was whether the offer should be made, with people deciding whether to accept it or not, with the offer increasing until staffing for the school is complete.

Well, beyond the morality of our question - Teachers (at least here in the NE) have very strong unions that will not allow that sort of "what will you take to risk your life approach".
And then there is the moral point of view that basically says it is not ethical to make that offer and make employment conditional on risking ones life.

True disclosure - Mrs TZD is a secondary school teacher and she and colleagues are not about to go back into the teaching environment that they had prior to February - and at least here in New York, the union can prevent it.

[CDu]

I mentioned in another post that the limited testing in the early stages of the disease is likely a factor in the phenomenon that cases are rising but deaths aren't at the same rate. Here's an illustration of how that could be playing out, courtesy of IHME. Spoiler alert: it might be playing even more of a role than I had thought.

In the first graph below I created a longitudinal plot of observed cases, deaths (you can barely see it), and "estimated actual" cases from February through the end of June (again, courtesy of IHME's data). The "estimated actual" cases assume a couple of things: in the early stages of the pandemic, we were missing cases by about tenfold (catching about 10% of actual cases), and that the cases which were identified were identified a few days later in the process than now. So as you can see, the estimated cases dwarf the observed cases in March and April. But, as April moves to May, and then June, testing increases substantially. As such, the disparity between observed and estimated "actual" cases decreases, both in terms of magnitude (catching more cases) and time of identification (catching them a bit sooner). By June, the multiplier from observed to "estimated actual" cases is only 2-3 times, and there is virtually no lag in identification.

COVID1.jpg

Seems reasonable enough, right? Plenty of folks (myself included) have suggested that we were probably underestimating cases by up to tenfold. And as testing has increased, presumably we'll miss a smaller and smaller proportion of cases.

The graph above is interesting, and also very concerning. Here's why. If we take the "estimated actual" cases and divide by 100 to create an approximate mortality rate assuming 1% mortality, and overlay that trend on the death data, here's what we get:

COVID2.jpg

Pretty striking, right? These estimates of the true cases suggest that the death rate is really about 1% (which is what most researchers think) and that deaths lag true cases by about 3 weeks. And if this is correct, then we're about to see an uptick in deaths again. Because the confirmed cases have spiked back up dramatically in the last 2 weeks (the data above stop at 6/30), and we've already seen the deaths starting to trend up a bit.

Now, this is all hypothetical, as the "estimated actual" are just estimates. But if they are close to correct, then that would suggest that the biggest driver of the trends in deaths really is the case load, and unfortunately we just severely underestimated the cases early on. And that the deaths are probably heading back up over 1,000 again in the very near future as we've probably had over 100,000 infections per day for the past week or so. And if the case trends in California, Texas, and Florida keep increasing, yikes.

I'll try to update this with the data from the last two weeks, and also add new data as it comes in.

[sagegrouse]

Total new cases today are 65.5 thousand, with a seven-day average of 62.2 thousand, which is nearly twice the peak average reached in April. New deaths are 465 for Monday, and the seven-day average has grown to 753. The new death seven-day average is one-third of the peak in April but 50 percent higher than a week ago (513).

The ratio of average new deaths to average new cases two weeks earlier was seven percent in April; now it is a bit less than two percent. The reasons for the difference, in my impression, are that we are doing a better job of protecting vulnerable populations like seniors in institutions; we are providing better treatment; and we may -- may -- have a more complete reporting of cases now. (CDu's post just above this one, which I just read, discusses likely mortality rates from the coronavirus.)

The worrisome thing is that, even at a two percent mortality, with new cases now averaging over 60 thousand a day, the daily number of new deaths will grow to exceed 1,000.

Eyeballing the numbers of new cases by state, Florida and Tennessee exceed 500 per one million population. Texas, Louisiana, South Carolina, Alabama and Kansas exceed 300.

The states with the most numbers of new cases are, again, Florida (12.6 thousand), Texas (9.2) and California (8.4) -- together constituting bearly one-half of the total number U.S. new cases (65.5).

[devildeac]

The hospital that employs me has added a second COVID unit of 12 beds to our first COVID unit of 16 beds which usually ran a census of 10-12 patients in April and May and 14-16 patients more recently. Uh-oh.

[OldPhiKap]

The hospital that employs me has added a second COVID unit of 12 beds to our first COVID unit of 16 beds which usually ran a census of 10-12 patients in April and May and 14-16 patients more recently. Uh-oh.

Not good, stay healthy. (And tell NCSisterPK the same if you see her!).

[devildeac]

Not good, stay healthy. (And tell NCSisterPK the same if you see her!).

Will do. Haven't seen/talked with her in a couple years I'd guess.

[Furniture]

I always check this site for my eye test data analysis.

https://covidtracking.com/data/us-daily

This is one way that I see it. Back in mid April we were running at about 50000 or so hospitalizations but deaths were about 2000 per day. Fast forward to the present and we are back around that same number of hospitalizations but average daily’s deaths are less than half. 800 or so ish...

Since presumably most of the deaths were in New York In April could it be that many died because they could not get good treatment? Lack of ventilators, space in ICU’s etc..Just a thought.

On another note hospitalizations in NC dropped for the second day running after about two weeks of increasing.

[budwom]

Miami-Dade ICUs reported at 96% of capacity, test positivity rate of 26% in the area. Not many ways to make that look good.

more from there:https://www.miamiherald.com/news/cor...244190287.html

[sagegrouse]

I always check this site for my eye test data analysis.

https://covidtracking.com/data/us-daily

This is one way that I see it. Back in mid April we were running at about 50000 or so hospitalizations but deaths were about 2000 per day. Fast forward to the present and we are back around that same number of hospitalizations but average daily’s deaths are less than half. 800 or so ish...

Since presumably most of the deaths were in New York In April could it be that many died because they could not get good treatment? Lack of ventilators, space in ICU’s etc..Just a thought.

On another note hospitalizations in NC dropped for the second day running after about two weeks of increasing.

I wonder, Furniture, if the high early death rate was the poor original condition of the patients and, since then, we have done a much better job of shielding those in nursing homes and assisted living from the coronavirus. Of course, COVID-19, early on, overwhelmed medical facilities in the Northeast, certainly affecting the standard of care. Also, one factor CDu cites is the "under-detection" of coronavirus infections.