Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
Day 74: COVID-19 Italy Has Turned the Corner
The data is starting to be compelling that Italy has reached its peak and the epidemic is beginning to recede:
Deaths are lagging indicators so that when you compare the slopes of the two graphs above they make out a distinct decline in COVID-19 cases in Italy. This, of course, does not mean the illness is over or that the deaths will not continue to mount. But it does mean that there will be less illness and death going forward.
The US, nationally, is about 14 days behind Italy. But our data is heavily skewed by the NYC area. Ironically, the peak for NYC is about 7-10 days away because they do not have a flat curve, while the peaks for other localities that have flattened the curve are some weeks away. Until people see that NYC is over its peak it will be difficult to have the real conversation: when do we let people go back to work?
Apropos of that, the latest scare is that the COVID-19 indeed can be aerosolized and that everyone should be wearing the best face-covering they can manage. But the reporting on this does not address the irony that this represents: If the virus is aerosolized then many more people are exposed/infected than cases confirmed. If many more are exposed/infected then the percentage of all persons suffering serious illness and death from COVID-19 is even smaller than currently envisioned. This changes the risk profile. It also means that simply having a comorbidity is not a death sentence as the numbers of persons with a comorbidity is very large.
That is not to say that the persons who are afflicted with a severe case of COVID-19 are not suffering greatly. There are also some number of persons (as yet undetermined) who will suffer from chronic pulmonary insufficiency even after recovery. There is no need to trivialize the severity of this disease. But from a public policy standpoint, the weighing of health harms between the disease and poverty becomes more skewed in favor of poverty just based on the numbers.
And that is where an excellent article by Craig Medred, an independent reporter out of Alaska comes in. I have referenced his reporting periodically. His latest piece is Fear fear. In it Craig outlines the evolving cost-benefit considerations of lockdowns versus other public health strategies:
“Lockdown is going to bankrupt all of us and our descendants and is unlikely at this point to slow or halt viral circulation as the genie is out of the bottle,” they wrote. “What the current situation boils down to is this: is economic meltdown a price worth paying to halt or delay what is already amongst us?”
The idea has gained some traction in the medical community, but not much. Political leaders, meanwhile, have largely gone in the opposite direction. The United Kingdom suggested it might let the virus spread enough to create what is known as “herd immunity,” but quickly backed away when some scientists and the public protested.
The Dutch suggested the same idea, backed away, but are now studying it. Meanwhile there is the suggestion from many scientists that herd immunity is in some way inevitable.
At that point, the disease becomes unable to easily jump from person to person and fades out. This is herd immunity. Unfortunately, some pathogens – most notably the flu – are able mutate and again return.
Katz and Heneghan have suggested that the best way to get herd immunity might be to shelter those vulnerable to fatal COVID-19 infections and let the disease run much like the flu in the rest of the population.
“The data from South Korea, where tracking the coronavirus has been by far the best to date, indicate that as much as 99 percent of active cases in the general population are ‘mild’ and do not require specific medical treatment,” Katz argued in his NYT op-ed. “The small percentage of cases that do require such services are highly concentrated among those age 60 and older, and further so the older people are.”
But responding to COVID-19 in this way at a population-level generally runs counter to the beliefs of Western societies that prize individuals. The mere possibility that a previously unknown disease could kill younger people – and it has – appears to terrify much of the Western world.
Thus Katz’s suggestion of an alternative approach aimed at protecting the elderly and those at risk because of ill health while putting everyone else back to work has to date gained no political support.
Whether it will ever gain serious consideration is an unknown, but there are more than a handful of scientists who share Katz’s concerns about long term problems inherent in the current strategy.
And so should we all. I am going to make a prediction: by Easter, New York will be past the peak. When that happens, but sadly not much before that, the President can entertain his initial instincts that opening up America for business again needs to happen sooner rather than later.
Italy has turned the corner. China (although its totals are suspect) has done so as well. Spain will turn the corner about the same time as New York. By Easter, most health officials will still be focusing on understanding the disease, but the picture will clearer that the health crisis is receding and the economic crisis has to take priority.
[Note: Links to all my CoVID-19 posts can be found here.]
Published in General
For those tracking things, the worldometer site has added a test count column to their country table, where available. Probably subject to the usual data quality issues, but interesting.
“…
And you, my father, there on the sad height,
Curse, bless, me with your fierce tears, I pray.
Do not go gentle into that good night.
Rage, rage against the dying of the light.”
– Dylan Thomas
The New York Times did a hate piece against. Bolsonaro on this issue, but didn’t claim that he called it a hoax, complete or partial. And he didn’t order the population to ignore all warnings. He is downplaying the threat, but he hasn’t gone that far.
There’s fairly good data from Germany that well over half of people who go onto ventilators end up surviving. Nobody can say how applicable this is to the US, but I’d argue more applicable than any figures from Spain, Italy, or China.
That being said, I agree with your general sentiment. In the two countries by far the largest Covid-19 epidemics (Spain and Italy), >80 year olds make up over half of the dead and most of them have multiple comorbidities, i.e., they were likely not long for this world in the first place. I also just read an article by the head of the Spanish sort-of CDC who said that many of the people dying of coronavirus would have died within the year anyways.
So even if most people that go on ventilators end up surviving, there are still groups of patients who are likely to die no matter what.
Now that reasonably robust data on deaths and other clinical outcomes is becoming available from countries like Germany, England, France, and Italy, we need to have a closer look at the demographic breakdowns and have a difficult conversation about when it’s worth prolonging a life.
We should also look into ways of providing ultra-high risk patients palliative care outside of a hospital setting but still in some other controlled setting so they are not simply “left to die”. And we should set up an advanced directive-type system allowing such risk candidates to clearly state their wishes in the event of Covid-19 infection (above and beyond the general instructions in an existing advanced directive, which some doctors may not honor as applicable).
I wouldn’t put too much hope in Brazil as a counter-test case. For one, many municipal leaders are apparently taking matters into their own hands and locking down their populations with or without Bolsonaro’s order. Second, any statistics from Brazil would be questionable either way. And given that lack of statistics, the only other measure will be anecdotal, and we know the press will play up the devastation to the max.
In that vein, I have seen numerous stories in different outlets about Ecuador over the past two days. Supposedly the epidemic is spiraling out of control there. Yet all of the stories seem to draw upon the exact same body of anecdotes: man-on-the-street accounts of lots of deaths, an undertaker saying he’s been disposing of 5x as many bodies per day as usual, and the exact same picture of a make-shift coffin on the sidewalk.
Is that evidence of an epidemic out of control? Or just the natural result when an area with a very rudimentary health care system sees a modest uptick in deaths? Nobody knows.
This gets back to the question of how many actual strains there are of COVID-19.
I’d bet that there are at least two major strains, possibly more. One is the “killer” version, found in central China, Iran, Italy, and parts of New York City (by way of direct infection from Italy and China). The others are what’s hitting most of the rest of the world, with higher infection rates but much lower fatalities.
It’s also possible that the milder strains were out in the population well before November of 2019, and the nasty one is a later mutation, only hitting the people who haven’t caught the less-lethal version. This would explain the Diamond Princess, which would have more people who traveled extensively before that cruise, and who would have a higher chance of being exposed to the weak virus.
Viral genomes are continuously being sequenced from patients around the world, and no major differences have shown up yet that correlate with the notion that the hard-hit areas have a different strain than less hard-hit areas. This is one of the few areas where we have pretty decent data, because it doesn’t take much work these days to sequence an entire viral genome.
The problem with that is that it doesn’t take a major difference in the genome to make a big difference in the disease.
There are multiple versions (some researchers claim they’re not different enough to be considered “strains,” as they basically have the same effects on humans), but it’s really not farfetched to think that a simple difference might make one of them more virulent than the others.
The biggest issue is that, while there are various people out there running their sequencing machines 24/7. they still have to get samples – and most of the samples come from the same, known, obvious areas. We don’t know if there’s an undetected larger mutation that’s not very dangerous – they’re not taking random sequences from people who never get sick enough to see a doctor.
Absolutely agree on the antibody test. On China, the referenced report is pretty skimpy of details about the “rebound” outbreak. Rebounds are bound to occur. The question is rate and extent. And for that we do not have data.
But are bounds rebound to occur?
That data will only be inferred the epidemic is over. When deaths from various non-COVID-19 diseases that we have been recording for years are totaled for 2020 and compared to other years, your suspicion will be confirmed if a significant dip in the average shows up.
Apparently the reward for a successful lockdown is a delayed peak and a longer lockdown.
That goes both ways. Take the influenza virus: there’s a famous mutation (termed PB2 E627K) that can almost single-handedly render many avian-only strains of influenza into human pathogens. That’s an example of what you’re talking about.
At the same time, there are hundreds of other mutations in influenza strains that are recorded every year that appear to have no effect whatsoever on the species tropism or virulence of the relevant strain. So one mutation CAN make a difference, but doesn’t HAVE to.
Luckily, the scientists are one step ahead of you – they’re well aware of the bias in looking only at samples from one type of patient and taking steps to counteract it.
Remember that countries like Germany and South Korea have done so much overtesting that they have lots of samples from asymptomatic patients – people who would have never guessed they had the virus if somebody from the local health office hadn’t called them up, told them they were a contact person, and required them to get tested. So yes, lots of sequences have also been read out from asymptomatic patients, as well as patients with mild, moderate, severe, and fatal diseases.
And disease severity is only one of the dozens of parameters taken into account when choosing which samples to sequence. Others include: geographic location, presumable geographic origin of the outbreak (i.e. China, Italy, etc.), different time points of infection within the same individual, different locations within the same individual, different locations at different time points, patient age, presumable route of infection……
And I’ll summarize the findings so far on the correlations of sequence with severity of disease outbreak: bupkis.
The most interesting finding to date, in my opinion, is that the virus replicating in the lungs may have a different sequence than virus replicating in the upper respiratory tract. It’s not clear whether that’s a causal mutation, although such a mechanism is known from some flu strains. Interestingly, though, it doesn’t appear that the mutated virus is transmitted – instead, the “original” viral sequence in the upper respiratory tract is the one that seems to get passed to others.
Germany is really taking the lead here. Several independent antibody studies have been rolled out over the past weeks, some of which are smaller-scale aiming to get some smidgen of data on the board as quickly as possible, and other longer-term projects aiming to test the same group of about 10,000 people on a regular basis to track the growth in seroconversion over time.
Either way, it looks like some key data may be coming within a few weeks.
That is very interesting.
Just a reminder: talk about rates of fatality, severe cases, and “turning the corner” are entirely dependent on knowing how many people are actually infect and how many were infected yesterday.
We have no idea.
Confirmed cases are mostly a measure of how many tests have been given and to whom.
Is there any working theory among virologists as to why young people under twenty years old are not getting sick from exposure to the virus?
Is there speculation that the reason is that they are able to develop antibodies quickly to it, so quickly that they don’t become sick from exposure to it?
Or are they, for some reason, not becoming ill from it because of some sort of innate resistance to it?
Or is the difference perhaps anatomical in the structure of their immature upper respiratory tract? For example, little kids get inner ear infections mostly because there is a crook in the ear canal that straightens as children get older. It’s in the crook in the canal (Eustachian tube, I believe) where secretions build up, creating a medium for bacteria and viruses to multiply.
What do you think, Mendel ( @mendel )?
it’s not true that we have no idea. We have a rough idea. How rough is hard to say.
And talk about turning the corner is also based on death rates, not just on number of confirmed cases. It’s a lagging indicator, but death rates seem to be telling the same story. There are problems with the data on death rates, too, but even though they aren’t recorded consistently in all jurisdictions, they are less rough than the data on confirmed cases.
Until the numbers start to rise again because everyone is out and about again. Then what happens? Will this first round serve as an inoculation of sorts and keep people from being willing to shut everything down again? Or will we find ourselves right back where we are?
You are describing the “rebound” effect. It’s going to happen, the only question is to what degree. If it is done right (a big assumption) the powers that be will see that the trend is sloping downward and will loosen the grip on the economy (but not release completely) with cautions to continue to observe hygiene and distance measures to prevent spread. They will also set expectations that there will be an uptick in new cases over the next couple of weeks after restrictions are eased but that it will not reverse the downward trend if people continue to observe hygiene and distance.
Italy’s data from yesterday is now included and (knock on wood) the downward trend is holding. Deaths were up slightly from yesterday but well below 4 of the previous 7 days. Daily cases peaked on March 21 at 6,557. Yesterday’s count (4,585) is less than all but two days since the peak.
Lessons (being) Learned: A very worthwhile post at ChicagoBoyz
Oregon Health Authority stops providing certain data about hospitalization because it doesn’t trust the data.
That all sounds nice. Now throw in the media’s response to the rising numbers – which I suspect will be very similar to what we’ve seen this time around. The numbers are going up again! The Apocalypse has resumed! We have to go back on lockdown if we want our parents and grandparents to survive!
Yes. Let me highlight 3:
Yeah, that’s a problem. After telling us to flatten the curve, they run headlines ignoring what is happening to the curve and focusing on the big numbers.
Data gathering in a crisis environment is very difficult. Seeing tabular summaries hides the variety of judgements that are being made by those providing input to the database. And it can be amazing how differently people interpret guidance for reporting.
Years ago I found my department in a crisis of sorts. Upper management was demanding data to be aggregated and reported in a fashion that the then current inventory system did not support. People with the best intentions were trying to kluge the data together through a series of sorts that were then aggregated into spreadsheets for dissemination. As I was looking through the daily reports I saw inconsistency in the numbers and only after talking to the responsible supervisor did I discover the process through which the reports were being developed. Problem #1 was that they were changing criteria almost daily without any notation of the changes. I put a stop to that immediately. Right or wrong the daily data was to be reported on a consistent basis. Changes to criteria were to be discussed and formally modified, including re-running historical data so that trends were mapped against the same criteria.
Data needs shift over time. Reporting systems prove inadequate to the new demands that were not contemplated when those systems were created.
Yeah, I’ve been meaning to tell my story of how that worked with scientific data in my discipline, but it’s hard to convey accurately just what was going on without putting more work into it than I’m inclined to do right now.
I agree that death rates are a better indicator (although lagging, as you say) of the trend, although, there are reasons the death rate might fall besides the overall infection rate falling. But I don’t think it’s true that we know how many people have been infected to any meaningful degree of precision.
Maybe it’s semantics. I’d say “how rough is hard to say” is doing more work than “we have a rough idea.” Do you disagree? If you were asked to give an estimate for the total number of infections to-date in the US, how wide would your range be?
In fact, in the spirit of Alex Tabarrok, let’s bet. You give a 50 percent confidence interval for the number of infections as of right now, and I’ll take either the “inside” or the “outside.” Loser sends the winner 12 oz of hand sanitizer…
That’s assuming the Brazilians actually believe him, and take 0 preparations.