Coronavirus - worldwide spread and control

It does not cease to surprise me how some people keep trying every sort of rationalisation they can think of, to believe that COVID19 is less devastating than official numbers report. The same sort of people on Twitter deny that longcovid is severely disabling, or insist it is simply due to anxiety etc.

 

I could hardly make this one up. Florida will not release data on COVID-19 in schools due to "privacy concerns". Similar concerns previously applied to infections in nursing homes and jails.

 

I think this group has a 'the ends justify the means' attitude and that generally speaking that attitude stems from a desire to create a world where everything is highly politicized.

I think too that they often confuse their highly developed abilities to rationalise with reason.

 

Wait did the UK govt’s app for contact tracing ever get off the ground? There was so much publicity around that?

(I’ve been out of the Covid loop for a while). Why don’t we hear about it anymore?!

And what’s happened to “Vaccine by September?”

 

they scrapped it after they 'found out' it was incompatible with a lot of smart phones.

 

There's a reason I've stopped giving daily reports on covid-19 cases in Florida. I started to question the numbers, because they didn't show internal consistency.

This week did nothing to improve my opinion, with over 75,000 old results suddenly found by Quest Diagnostics. The state is blaming Quest for dumping results, and making us look bad, and the Governor has cut off state contracts to Quest. I'm wondering why no one noticed that many results missing. I've wanted to know how many test results were pending for months. Apparently, the people running the state operation didn't know either.

I've already mentioned adviser Scott Atlas, M.D., who favors spreading the virus among those at low risk. He says he favors protecting the vulnerable, but there are three problems with this: 1) visitors to elder-care facilities will not be tested if they don't show symptoms; 2) asymptomatic school children will not be tested; 3) asymptomatic restaurant workers will not be tested.

What I've laid out above is a range of pathways to infect those in long-term care facilities, and those outside them, like diabetics, who are also at increased risk.

Reputable researchers believe that roughly half the spread is by people who are asymptomatic, presymptomatic or oligosymptomatic.

What I can glean from data being released is that we are running over 3,000 new cases per day and over 100 newly reported deaths per day. This is a long way from "not a problem", and a prediction of 300,000 U.S. deaths by year end seems reasonable. Florida alone could contribute tens of thousands beyond the 11,903 already reported.

Added: today we added 3,571 cases and 149 deaths. Three more non-residents died, so we should be reporting 152 newly-reported deaths. Cumulative total 12,055 deaths. We could catch up with California. This is not a small problem.

Added later: those were Friday's numbers above. Saturday's are 3,656 cases and 61 deaths, counting only residents.

Another critical metric is the percentage of newly tested people testing positive. This is where I can't make heads or tails of what is going on. If I just take the raw number of people who have never been tested, and divide this into the number of new positives, I get 14%. Of course, some of those testing positive may have previously had a negative test. Adjusting data for this doesn't seem difficult.

Aside: what I do not see is any effort to take random samples to gauge how well testing is working. Nor do I see efforts to find hot spots.

Other estimates of the percentage positive range from 22.779% to 6.22%, with official sources favoring lower numbers. Anything above 10% is a "red flag" warning for epidemiology.

Official efforts to hide or delay data, or downplay the problem, have had precisely the opposite effect on me. I'm afraid we are flying blind into a bigger disaster.

So, here we are with schools open, and a labor day long weekend, both of which I expect to produce a surge in cases at the time when flu season begins.

BTW: the administrator who has been leading ACHA, the official responding to the crisis, has just resigned to take a lobbying job.

 

Do you know how many people die from all causes in Florida at this time of year under normal circumstances? If you don't then you have no context for the over 100 newly reported deaths per day, and may be scaring yourself unnecessarily.

According to Worldometers there have been 51 deaths attributed to Covid-19 in the whole of the UK in the most recent 7 days they have figures for. However, according to the UK government's statistics department, the average number of deaths in just England & Wales (i.e. excluding Scotland and Northern Ireland) for the week ending 21 August (the most recent week for which data is available) over the last five years was 9,157. So the Covid-19 deaths are just a drop in the ocean compared to all deaths that would normally occur at this time of year.

 

I think my point still stands though. If people don't know how many people die under normal circumstances in an ordinary year, then covid-19 death numbers are numbers without context, however they are recorded.

 

Here's an account of the problem back in April. For that month Florida typically had around 17,000 deaths, this April we had 19,200, say 2,000 excess deaths. By official counting only about half of those were COVID-19 deaths. State criteria, though different from the U.K.s, require several things that make it difficult to catch all deaths caused by COVID-19. People could easily die without being tested. Epidemiologists have assumed most of those excess deaths were due to the virus. This gave a much better prediction of what came next than the state arguments for reopening, which violated strict CDC recommendations, unless these were watered down.

Note that 100 deaths per day adds up to 3,000 in a month or 36,500 in a year. These are not negligible numbers. People at increased risk may have three times the personal risk as aggregate numbers suggest.

Aside: you could take the attitude of one teenage friend who responded to my reading about an earlier pandemic. After thinking about the horrifying numbers I had mentioned, he said "All them people would of died anyway." As one who will ultimately die, I don't recommend this as public health policy.

Here's a NYT article on excess deaths in many states.

Florida also had a dispute between local medical examiners who listed COVID-19 as the cause of death and state officials back at the beginning of the pandemic. The response was to block medical examiners from releasing their own numbers. This does not reassure me.

In most epidemics you have a category of probable cases, even if you do not have a test with 100% sensitivity and selectivity. If you can find such official numbers, I'd like to see them.

Here's another analysis specific to Florida which also suggests substantial undercounts, though not the extremes some have suggested. U.S. statistics do not show the kind of bulge in excess pneumonia deaths seen in other countries. This does not discount the possibility some were infected in late December 2019. Where pathologists have kept samples from puzzling cases we may ultimately push the U.S. start of the pandemic back into 2019. It is already traced back to January.

Florida officials have repeatedly downplayed the role of asymptomatic infected, and have implemented policies which ignore the possibility. (Check what I've said above about restaurants, schools and long-term care.) This is what concerns me in terms of epidemiology. At present community spread is so common it reduces the value of contact tracing, for which we never had an effort commensurate with the problem.

In terms of regaining control of the pandemic, I have to fight despair. The measures required at present would be too extreme for public acceptance. Predictions of 300,000 U.S. deaths by the end of the year, or 400,000 by the end of January look far too plausible.

 

In the US, much of this unreliable covid data nonsense is part of massive organized disinformation and astroturfing campaigns designed to divide, distract, obfuscate and cast doubt in an election year.

I'll leave it at that.

 

I think this recent blog post from Malcolm Kendrick deserves an airing in this thread :

COVID – why terminology really, really matters

I found the post well worth reading.

 

I do not see that this is very relevant. The same thing could be said about a motorway pile up because a lorry drove the wrong way but it would not be considered a drop in the ocean compared to all deaths.

Deaths from covid are down because the ones most at risk of dying are still taking maximum precautions to prevent them. Younger age groups now make up the majority of new infections so deaths may not be so relevant but longcovid/ ME may be.

Nothing has changed since March, the virus is still out there and given the chance it will run rampant again. The death rate may be lower but only because many of the most vulnerable have already died.

 

It seems to me a remarkably bigoted and stupid analysis to be honest.
He just wants to pretend there isn't a problem.
Ask my niece who works in A?E if there is a problem.
The problem is that there was no initial closure of air traffic that would have aborted the entire European epidemic. As it is the failure to act has cost hundreds of billions in lost revenue in addition to tens of thousands of premature deaths.

 

I suspect that I would largely agree with Jonathan Edwards post above, but I haven't read that article. I was busy reading other things, and I only have one pair of eyes. I'm not going to get into a flame war if I can help it. If the author is arguing for a distinction between case mortality rate and infection mortality rate, I can agree, while pointing out we still have a poor idea of infection fatality rates. We have to go with information available. Mortality rates for those admitted to hospital are improving, but they are still nothing to brag about. Florida alone has thousands in hospitals for COVID-19 today.

My viewpoint from the outset as been borrowed from epidemiology. That's something with applied mathematics I can evaluate. What has become completely lost in a great deal of discussion is the purpose of epidemiology.

The purpose of epidemiology is not to take political power or assign blame, it is to find interventions that reduce the ultimate burden of morbidity and mortality caused by an epidemic. It doesn't matter if you don't start out without any available interventions, as Dr. John Snow did. When his "ghost map" led him to a particular pump providing contaminated drinking water, locking that pump provided an intervention no one had previously tried. (There is a great deal more to that story, and it wasn't quite the instant triumph students hear about. Nobody in Victorian England wanted to believe the fecal-oral route of infection. The pump was returned to service.)

Plotting cases is vital, even if you don't have a fantastic test with scientific sophistication you can operate on the basis of symptoms and clustering. Waiting to plot deaths is not the best way to do this. In the case of AIDS there might be 10 years between infection and death. It took a while for people to believe the lethality of the untreated disease. Tuberculosis might not kill for decades, and there was a great deal of bigoted opinion about genetic inferiority, even while upper classes were sending their infected relatives off to places in sunny climes or on mountaintops to die where the news could be limited.

Finding hot spots is vital for identifying transmission. So is finding locations or subpopulations without cases. These suggest possible interventions. During the 1918 "Spanish Flu" cities that reduced public gatherings and used masks did a great deal better than those that did not.

Next, the purpose of tracing contacts is not punishment. Even if you find a classic "Typhoid Mary" you isolate and treat her, instead of punishing her for deaths she unwittingly caused. People have tried other strategies, and guess what? Punishing the infected doesn't stop the epidemic, it just makes it harder to track, and encourages the spread by causing infected people to flee. You have to provide some positive incentives to help the infected.

I'm going to take a small exception to Jonathan's post to illustrate the way you can handle imperfect tests and treatments to stop an epidemic. South Korea sent flights to repatriate Korean nationals in Italy. They checked them for symptoms before the flight, and excluded those with active infections as best they could. They required masks on the flight anyway, and some people became infected anyway. This did not create a new outbreak in Korea, because those evacuated this way were quarantined when they arrived. Those quarantined were regularly monitored, and treated when they showed symptoms.

General principle: imperfect solutions require more burdensome interventions. Many governments failed to do much of anything effective because they were busy denying the problem, and the interventions were expensive. South Korea, Taiwan, Iceland and New Zealand brought infections under control. If they can keep things under control for another six months, there will be new interventions, some much more effective.

Nothing dooms to untimely death those who are currently vulnerable to SARS-CoV-2.

 

I think something similar is going on in Sweden - the statistics are being filtered and released occasionally in batches.

 

Research suggests that testing may also be detecting "dead" virus particles:

http://www.msn.com/en-gb/health/med...h-suggests/ar-BB18JRoa?li=AAnZ9Ug&ocid=ASUDHP

 

My logic may be off but do most people not get tested when they develop symptoms? Unless there is a long delay before testing they will still be actively ill.

I am not saying the findings are wrong, but they only seem relevant in those cases where everyone at a school or factory is tested when some people are shown to be ill with the virus. But in that situation, some people will test negative because they have not be infected long enough to test positive so both false negative and false positive tests will be found.

Local lockdowns and quarantines still seem the best solution to stop the virus spreading.

 

I'm aware of the possibility of detecting viral fragments which are not infective. The most likely source for these is an active infection in a part of the body that is hard to sample. We already have examples of active individual infections which last for months.

Another problem is that when you find clusters of positive tests you are looking at the footprints of transmission. It may have happened in the past, but it is the only lead you have on current transmission. If you don't follow such leads and interrupt transmission the cluster will spread.

The problem here is not just the probability of error, but the consequences of the different kinds of error. Missing an active chain of transmission, or failing to break it, can kill people who were not initially involved in testing or quarantines.

Epidemiology always works with incomplete and ambiguous information, and waiting for ironclad evidence to take action that inconveniences people means allowing the epidemic to burn itself out, at whatever costs to those infected.

A long time ago my fellow citizens decided it was vital for me to spend two years of my life fighting a war, with an option to lose everything. Despite misgivings, I served. The sacrifices people are being asked to make to save hundreds of thousands of fellow citizens are not in the same league.