Added 12 march:
We had a bit of a conondrum when a legit looking analysis appeared in the Daily Mail. I normally won’t even look at the Daily Mail since it is a rag of teh worst kind. But, commenter Joseph M., a long time trusted friend and VSP (very smart person), and a scientist, dug in a bit farther, and he makes a convincing argument that this is worth posting.
The Daily Mail piece is here.
The following are Joseph M.’s notes and comments on it, and some graphics:
Professor Mark Handley, absolutely checks out – he’s much more than legit. As are every one of the other sources.
So I said I’d check out this article and I did just that, vetting it thoroughly. (Several hours of investigative journalism on a work of (initially) questionable journalism.) It’s the real deal, all right – it’s just a bit unfortunate that the journalistically despised tabloid The Daily Mail got the story instead of a respected London newspaper like The Guardian or The Observer. But WTF – that’s just the way this particular cookie crumbled, and one has to give credit where credit is due. I think they scored a scoop, and that were it published in one of the aforementioned establishment papers, world news media would have picked it up and run with it. (Maybe they did, but I didn’t see anything. As I may have told you, my friend Allan sent me the link – and I blew him off because of the tabloid source.)
The inconvenient truth is this: All the experts cited and interviewed in the article are at the top of their respective professional games. No slouches, these guys. The only question is, how much credence do we want to give to a computer scientist – are doctors not echoing his projection because they’re inherently more conservative, or because they have legitimate reasons to question his methodology – or are they falling back on a lame argument from [their medical] authority? I just don’t know, but I also know – we both know – that government reassurances and Anthony Faucci’s public cautiousness notwithstanding, the numbers we’re seeing exploding all over the map seem consistent with Handley’s argument. (Of course, were this my field of expertise and I had the patience to plug in all the numbers, I’d create my own graphs just for the sake of comparison. For whereas this was an off-the-cuff tweet by a concerned professor who discovered an alarming pattern – not a formal paper with any stated methodology – it looks like simple Algebra 101 to me – just numbers of cases as a function of time for each country … and compare the slopes, allowing for the slight offsets in commencement of infection …)
Mark Handley, the guy who created and tweeted the attention-grabbing graph, is Professor of Networked Systems, University College London; a member of The Royal Academy, and (a quick web search confirms) highly regarded by the epidemiologists and infectious disease people with whom he consults. He’s a computer network nerd – what can I say?
We are, after all, dealing with nothing if not network theory – and obviously his kind are crucial to epidemiological modeling; as we both realize, they work closely with I.D. specialists everywhere.
There’s also a prominent Chinese computer scientist and systems engineer, “Eric” Feiqi Deng, Professor and Director of the Systems Engineering Institute, School of Automation Science and Engineering (South China University of Technology, Guangzhou) – who sounded alarms early on and also tweeted graphs, like the simulation of Covid-19 transmission scenarios I sent you a day ago (and am re-attaching) – I can’t remember if you (or I) already uploaded it.
Of course, there’s been a veritable deluge of similar computer simulations and graphs – nothing new here – except for the interesting part, namely that ”we must flatten the curve” – terminology I’ve heard Trump administration people, CDC officials, and even reporters frequently use – is quickly becoming the newest American meme. [My physician friend Allan Wang (who himself has a deep understanding of infectious disease dynamics) forwarded me Deng’s tweeted (or published?) graph – and I’m unable to locate it online for context or proper citation. Don’t ask me why, but Deng took a helluva lot of international (and ad-hominem) heat for circulating this (I’m also guilty of dissing him in some emails, and I can’t even recall why!), and if I remember correctly he was slammed by (among other people) Harvard Chan School of Public Health epidemiologists – presumably for scaring the shit out of people when the graph went viral and for not being a biologist or physician.
The stink this raised lingeref for quite a few days – Deng was even interviewed in some newspaper articles – and now [this, alas, has become my constant refrain] I can’t even reconstruct what I read. Maybe my mention of the brouhaha will ring a bell with you?. Even without any context – the discussion (journal paper or tweet this was certainly part of), Deng’s graph is self-explanatory. – it’s essentially an elaborated version of the colored graph you posted on your Facebook page, or that article w/animated graph by the New Zealander that I posted there. (Frankly, this total immersion in Covid-19, compelling as it is, has me flailing around w/respect to my real obligations. What to do? The situation isn’t merely fluid or extremely dynamic – it might conceivably blow up into the worst domestic disaster we’ve ever seen – and no doubt the most unnecessary one, n the sense that, well, if only cooler heads were around to prevail … Did you read the anecdotes I posted on your blog about the violently irrational blowback I got from my brothers – all because of coronavirus?!)
Handley’s blunt tweet that “Everyone else will be Italy in 9-14 days time” was seconded by John K. Crane MD, PhD, Professor of Medicine, and Adjunct Professor of Microbiology and Immunology and of Pharmacology and Toxicology at the Jacobs School of Medicine and Biomedical Sciences, University of Buffalo – some obscure city in an obscure state I never heard of. http://medicine.buffalo.edu/content/medicine/faculty/profile.html?ubit=jcrane
So in other words this Daily Mail article (despite its clickbait bold black headline and equally bold bullet points underneath) – with its most compelling content consisting of unedited tweets (!) – is not conventional journalism, to say the least Still, it is solidly reported and chock full of valuable resources (compelling color photos; American news videos). I highly recommend that at the minimum we post the graph and its legend, and maybe include one or two screen captures from the more substantive tweets, with a link to the actual article.
The extensive Twitter commentary is informative, especially the tweets from Dr. Nick Christakis (see below). Your perhaps skeptical readers (perhaps pacified with a line or two from one of us to soften them up) will just have “get over” their reflexive revulsion from seeing The Daily Mail masthead. This is most certainly not the typical sensationalized puff piece on the latest shenanigans of some obscure fourth cousin to the Queen.
To sum up, Mark Handley’s alarming numerical projections are consistent with all the data we’ve seen – and (sadly) with the high likelihood that our government – this particular administration – will continue to drop the ball in some fashion or another and make things much worse, even dire. (DJT has some pretty-fucking-scary emergency powers at his disposal, and we both know how and why he’ll be inclined – and by whom prodded – to deploy them.)
Handley’s graph shows that rates of increase in 8 of the 9 countries he examined follow the same slope, albeit with a predictable time lag correlated with the date of the respective initial outbreaks. (The graph is pretty ‘busy” and somewhat hard to read (precisely because eight of the plotted countries follow identical, overlapping trajectories), but per the black color coding it seems that the single low-slope country is Japan, – which as we know caught’ Covid-19 from Chinese travelers early on, and so is probably starting to flatten out.)
The article is 50% wheat, 50% chaff. For example, there’s the stark Daily Mail headline “America will be in lockdown like Italy in less than two weeks” and similar (but remarkably restrained) editorializing … What I therefore attempted was to “migrate” the good parts to a Word document as a preparatory step for GLB / FB posting. But it was just too cumbersome – my draft Word “repository’ wound up being over 20 pages long! This is largely because of the massive number of follow-up tweets from very relevant players – especially Nicholas Christakis https://eeb.yale.edu/people/faculty-affiliated/nicholas-christakis and Jason Van Schoor, an anesthetist and clinical fellow at University College London https://twitter.com/jasonvanschoor?lang=en , evidently highly respected,* who at the end of this long Daily Mail article relays powerfully disturbing real-time reporting from his medical friends on the front lines in Lombardy. We’ve all see news videos to the same effect.
[ * https://virginia.sportswar.com/mid/13441707/board/general/ ? “I do not know van Schoor but he has had a Twitter account since 2012, has more than 8,000 followers which include some people in health care I know and respect. He was quoted today in an article by UK’s 3rd largest newspaper, the Daily Mail (link below). The fact that others are picking it up too does not make it sketchy.”]
Added 11 March:
From this source:
COVID-19 can be spread before it causes symptoms, when it produces symptoms like those of the common cold, and as many as 12 days after recovery…
…Researchers at Johns Hopkins found a median incubation period for COVID-19 of 5.1 days—similar to that of severe acute respiratory syndrome (SARS).
… novel coronavirus quickly begins producing high viral loads, sheds efficiently, and grows well in the upper respiratory tract (nose, mouth, nasal cavity, and throat).
“Shedding of viral RNA from sputum outlasted the end of symptoms,” the authors wrote. “These findings suggest adjustments of current case definitions and re-evaluation of the prospects of outbreak containment.”
… “In SARS, it took 7 to 10 days after onset until peak RNA concentrations (of up to 5×105 copies per swab) were reached In the present study, peak concentrations were reached before day 5, and were more than 1,000 times higher.”
Michael Osterholm, PhD, MPH, director of the Center for Infectious Disease Research and Policy at the University of Minnesota, which publishes CIDRAP News, said that the results challenge the World Health Organization’s assertion that COVID-19 can be contained.
The findings confirm that COVID-19 is spread simply through breathing, even without coughing, he said. They also challenge the idea that contact with contaminated surfaces is a primary means of spread, Osterholm said.
“Don’t forget about hand washing, but at the same time we’ve got to get people to understand that if you don’t want to get infected, you can’t be in crowds,” he said. “Social distancing is the most effective tool we have right now.”
…researchers estimated the median incubation period at 5.1 days (95% confidence interval [CI], 4.5 to 5.8 days). They found that 97.5% of patients who have symptoms do so within 11.5 days of infection (CI, 8.2 to 15.6 days).
After the recommended 14-day quarantine or active monitoring period, “it is highly unlikely that further symptomatic infections would be undetected among high-risk persons. However, substantial uncertainty remains in the classification of persons as being at ‘high,’ ‘medium,’ or ‘low’ risk for being symptomatic, and this method does not consider the role of asymptomatic infection.”
“The current recommendation of 14 days for active monitoring or quarantine is reasonable, although with that period some cases would be missed over the long-term.”
The sources of these comments:
Virological assessment of hospitalized cases of coronavirus disease 2019
The Incubation Period of Coronavirus Disease 2019 (COVID-19) From Publicly Reported Confirmed Cases: Estimation and Application
I’m not really going to live blog this virus, but I wanted to get a few thoughts down, and expect to be interrupted by a scheduled event in a few minutes. So, I’ll come back with more later. Perhaps this post will become a regularly updated/edited thing, we’ll see.
Call it 2019-nCoV (pronounced “Encovee”? — rhymes with anchovy –) because if you call it “coronavirus” you will have to spend time in the obligitory sidetrack: “Don’t panic about this virus, there are many kinds of coronaviruses, most of them are harmless, the common cold is a coronavirus.” That is much like saying “don’t panic about this serial killer that just started operating in your neighborhood, they are just humans, and most humans are totally harmless.”
Speaking of coronaviruses, yes, they are common, and this is likely to cause some, maybe much, variation in immunological response to Encovee, since there could be some cross effects of immunity from previous forms of the virus.
It is being noted by many that the flu is a much more common and deadly disease. Let’s talk about that for a second. Yes, it is, but most influenza viruses are moving across an experienced landscape of hosts that have a combination of prior immunity and vaccination. Encovee is treading on immunologically virgin ground. This likely means it will spread fast, almost with impunity. After that, maybe it will become just another one of the coronaviruses.
We really have no idea whatsoever what the rate of illness or mortality is. We can talk about this later, but this is a very complex and generally poorly understood thing. What we do know is that most people who get Encovee don’t die from it. We have no idea how many people are infected but show no symptoms, or the ratio of people who get a little sick vs. very sick, or, really, the ratio of those who get it and die. Graphs of the rate of its spread show an alarming verticality, but with mortality being a low almost flat line, at a very low percentage.
As of last night, here is what WHO was saying:
Their most recent situation report (of Jan 25) shows 1,320 confirmed cases, with most from China, HOng Kong, Macau, and Taipei. There were 23 confirmed cases outside of that area, 21 of which had history travelling to Wuhan, the Chinese epicenter. The others appear to be human to human contact within a family or similar.
Of a subset of 1287 cawses, 237 are counted as severe. There had been 41 deaths.
Note that all the scary numbers and charts you’ve seen, if you’ve seen them, are projections based on various models.
Projecting a disease outbreak at the beginning is like taking a bead on a certain direction and walking that way, and seeing where you get, but with this caveat: At the start of your journey, your compass sucks, and you don’t know how badly it works. Slowly over time, it improves, and it is hard at first to tell how much it improves. Eventually it starts to become a pretty good, but still limited, tool. Put another way, we can model the course (spread, magnitude) of a disease outbreak very very accurately — after it has happened.