The disconnect between Information and News (COVID-19 update, 26 Feb 20)

Yesterday morning the US Centers for Disease Control (CDC) gave a media briefing, led by Dr. Nancy Messonnier, Director of the National Center for Immunization and Respiratory Diseases.  The transcript still isn’t up, but you can listen to the audio here if you have 30 minutes.  After it was over, I was thinking “OK, nothing here to get excited over, not really any new data.”  Then the news reports started coming out and I started getting messages, headlines like this NBC article: “Americans should prepare for coronavirus crisis in U.S, CDC Says” and pull quotes like “It’s not so much a question of if this will happen anymore but rather more a question of exactly when,” an official said.   I thought “wow, did I miss something?”  so I listened again.  Nope.  So what’s going on?  Is it time to panic, get masks, and start drinking?

Maybe not.  Let’s start with the pull quote.  What Dr. Messonnier said in full was “It’s not so much a question of if this will happen anymore but rather more a question of exactly when this will happen and how many people in this country will have severe illness.”   In my view that puts a  whole different spin on things – the “how many people will have severe illness” is the key.  Certainly we should prepare – especially health care providers need to dust off their pandemic response plans (and, recall that Pandemic doesn’t mean people dropping dead in the streets, all it means is widespread).  Schools and businesses should probably think about what they would do if – and at this point, it’s a very big if with decreasing probability – things get bad.  Check ready.gov for some good advice and checklists.  Most of these things are stuff you should have already done to prepare for earthquakes, hurricanes, etc.

This points up a huge problem with how complex, scientific topics are covered by the media.  The details matter.  You can’t just grab a pull quote (and especially cut off the quote halfway for a headline!) and covey the nuance people need.  It stokes fear and uncertainty – and that has consequences.  Just look at what the stock markets are doing (some of which is justified because of Chinese supply chain disruptions, but risk taking on a life of their own).

Another problem is that Dr. Messonnier was discussing a lot of “what if” scenarios that got reported as active plans.  Things like closing schools, businesses, etc. are certainly things that would be done in the event of a bad contagious outbreak.  That kind of control measure can do a lot to limit the spread of a disease.  But the economic impacts are severe – and those actions have consequences that results in people dying.  Doing nothing can cause misery and kill, doing too much can cause misery and kill.  It’s a hard balance to keep, and is far more difficult for decision makers to keep when there is bad reporting that scares the population and the political class, and stampedes them in to making bad decisions based on fear, not facts.

The terminology can be a problem in science, and specialties matter a lot.  Here’s an example.  A couple weeks ago I talked to a Doctor who works for one of the agencies responsible for looking at COVID-19.  He said they were seeing mortality rates of 10% or so.  That would be catastrophic.  In looking over the data yesterday, I was calculating a mortality rate for those infected (to plug in to a financial impact model) at well under 1/100th of that.  So I checked in again, and he still said 10%.  In talking it through I figured it out – he was talking about cases with respiratory involvement, which makes sense because as a Pulmonologist  he only cares about those patients.  So when you hear things like mortality rates, the first thing to ask is “based on what denominator – total population, those infected, symptomatic, hospitalized, or what?”  That’s the kind of question and perspective reporters should be asking, but aren’t.

So what’s the bottom line right now?  Don’t misunderstand this – there is still some uncertainly around COVID-19, and as noted above some precautions such as those currently recommended by CDC  are sensible, but the numbers are starting to line up (link to PDF of summary spreadsheet) along the lines of a severe influenza strain, something like the 2017-2018 flu season. In that, the population mortality rate (fraction of the total population who died), taking in to account vaccination, is around 0.03 percent, the fraction of those who became symptomatic (got sick) who later died was around .14%, those hospitalized who died between 5 and 8%.

COVID-19 is almost certainly already here and spreading, we just don’t know it yet.  Testing is very limited both here and elsewhere.  And normal cold and flu strains, which are still going around and likely masking some COVID-19 cases,  cause a lot of death and economic impacts.  But in perspective, it is shaping up that COVID-19 is in that realm of impact, and is not the devastating pandemic that will like hit us someday if we aren’t careful.  In that sense, it is a good wakeup call (as if SARS and other outbreaks weren’t enough of a warning) and maybe we will listen this time.

Sources of information on COVID-19 (and bad reporting by the NYT)

First, here’s the link to the current US Centers for Disease Control (CDC) situation reports.   As the National Hurricane Center is the official word on hurricanes, the CDC is the official source for information on diseases.  The DHS/FEMA “ready.gov” site has some good advice on disaster planning, including pandemics.

Please be careful of other sources – even the major media (and especially bloggers – yes, even well connected and informed sites like this one) – and be careful of how CDC data is interpreted.  Determining risk in a dynamic situation like this is tricky.  For example …  from the New York Times today:

Experts have said that, based on the number of dead, the total number of cases in Iran is probably much higher, as the illness linked to the virus appears to kill about one of every 50 people infected.

I’d be really interested in hearing exactly what was said to the reporter(s).  There is no doubt that with 14 deaths, based on what we know, there would be a lot of infections.  But the statement that “the virus appears to kill about one of every 50 people infected” is extremely misleading.  A more accurate statement is that it appears to kill about one in 50 people who were able or sick enough to seek medical care and were diagnosed as having COVID-19.  The present state of understanding of COVID-19 (as of Monday, 24 Feb 2020) is that it seems like a lot of people get infected and don’t need to seek out medical care.  There is also uncertainty in the mortality in that especially in China it is likely deaths are under-reported.  But using the data from outside China, 1 in 50 of people who were sick enough to be checked seems reasonable.  So what does than mean?  Let’s say that conservatively only 1 in 2 people seek care – the other half think they just had normal flu or a cold.  That would make the rate 1 in 100 (1 percent).  But the likely number is more like only 1 in 10 people get sick enough to seek medical care.  That makes the rate 0.2 percent.  So you can see how sensitive the numbers are to figuring out how many people were actually exposed, and then got sick.

Exposed, infected, mildly symptomatic, sick, hospitalized, and/or died.  All these categories are important, and have to be carefully assessed.  If you are exposed, how likely are you to you to become infected?  If infected, how bad will it be, how much medical care will you need (if any), and so forth?  At this stage we don’t have great data (thanks, China) to get firm handle on some of these numbers – which means reporters have to take extra care to avoid either underestimating or, as seems to be the case here, overestimating, the real risk.

Addendum:  The Diamond Princess was a pretty good test case to see what is going on independent of Chinese stats. Out of 3700 passengers and crew, 634 people have to date tested positive, 328 are to date asymptomatic (not sick), 306 are sick to one degree or the other. 3 have died. So assuming those numbers hold, that’s about about a 10% morbidity rate (got sick) and 0.1% mortality rate (died), in a pretty bad environment for infection and spreading (most of the 3700 probably were exposed to one degree or another), but with good health care.  Usual disclaimers, small sample size, blah, blah – especially for mortality.   But morbidity is the interesting stat from my perspective, and one that says this isn’t so bad – from a medical standpoint. Perspective: the 2017-2018 influenza season, which was the worst in the last decade or so, in the US the morbidity rate was 13.8% of the total population, 6% had some interaction with a health care provider. 0.25% required hospitalization, and 0.018% mortality rate.

Additional note: To a large extent we are playing a guessing game with the numbers, especially the mortality rate.  People are extrapolating mortality rates by multiplying some rate based on reported cases (which drastically underestimates exposure) times the US population and getting numbers like 27 million sick and 270 thousand deaths.  But as noted, the Diamond Princess numbers are probably an extreme case with respect to infection rates; even a bit more isolation (as in a typical US city with lots of single family homes, and more space per person, etc.) are likely to have smaller rates.

Some perspective on the COVID-19 coronavirus and global response

Over the weekend there have been some major, and on the surface rather frightening developments surrounding the corornavirus outbreak that started in China last year.  So how bad is it, and should you be freaking out?  When the media starts using words like “pandemic” people start to panic.  But the bottom line on this situation is there is more reason to worry about the panic than the pandemic.  Here’s some perspective:

Coronavirus disease 2019 (COVID-19) is an infectious disease caused by SARS-CoV-2.  In English, that’s Severe Acute Respiratory Syndrome-related (SARS) CoronaVirus.  Here’s what the beast looks like …

CDC/ Alissa Eckert, MS; Dan Higgins, MAM – This media comes from the Centers for Disease Control and Prevention’s Public Health Image Library (PHIL), with identification number #23312.

Part of the problem is that the info coming out of China is pretty obviously crap.  We just don’t have a good picture of things like how contagious it is, and in particular nobody really believes the statistics on the number of cases and associated mortality rate.  That has given rise to lots of crazy rumors and fear mongering. But it seems like the following is true of COVID-19, based on reliable sources:

  1. It’s pretty contagious, and apparently can live outside a host for longer than stuff like the flu.  That means it’s easier to spread indirectly (someone touches a surface and leaves some virus, somebody comes along a day or so later and picks it up);
  2. A lot of people seem to just get mild cold/flu like symptoms – often so mild they are don’t realize they are sick with it, but are shedding virus everywhere;
  3. “Incubation” times are therefore suspect, and the 14 day time frame is problematic – some suggest up to 28 days;
  4. Some people do get pretty sick from it. It’s not completely clear yet just why, but aside from people who are normally vulnerable to respiratory infections (young children, the elderly and immune compromised), it seems smokers are more susceptible to getting really sick from COVID-19.  If it triggers pneumonia it gets bad (duh);
  5. With treatment/support care, it’s very manageable – just not in a crowded and populated megacity like Wuhan, etc. where it can spread quickly and stress health care systems.
  6. The test kits are limited in number, and may have a high false negative rate (in other words, say you don’t have it when in fact you have been infected).

Beware the term “pandemic.”  It conjures up visions of the Spanish Flu, the Black Death, and Monty Python Skits.  But all that word means is, per the World Health Organization, A pandemic is the worldwide spread of a new disease.  It really doesn’t say how bad it will be – just the flu, or something a lot worse.  COVID-19 is at the low end of that spectrum.  It seems like the mortality rate (the fraction of people who will die if they get it) is low, most likely less than the normal seasonal flu.  So, more contagious than the flu (bad), but perhaps less deadly (good).  Which means if you are reasonably healthy and take normal precautions (hand washing, don’t touch your face if you can avoid it, etc.) you probably won’t get it and if you do, it probably won’t be bad.  If you are in a vulnerable population (elderly, are immuno-compromised, or have lung problems) or are a moron smoker, be more careful and seek medical attention if you start to feel cold or flu like symptoms.

That said, the reaction to the virus will very likely cause far more damage than the virus itself.  Economic indicators are already trending sharply down on news from Italy that there are several outbreaks there, with authorities (IMNSHO overreacting) by putting in place quarantines.  This, combined with the virtual shut-down of production of many vital global exports from China (including a lot of medical supplies we have stupidly outsourced there just to save a couple bucks), means a global recession is inevitable at this point.  When it is announced that there are hundreds if not thousands of cases in the US, people will freak out even though as noted above there seems to be no more reason to freak out over COVID-19 than over the normal influenza virus.  Be prepared for a very rocky couple of months as authorities try to find a balance between trying to look like they are doing something, and causing more damage than the disease.  Once spring sets in (likely early this year according to both science and rodent based forecasts), it’s likely that like the seasonal flu, the infection rates will drop.  But by then the economic damage will have been done.

Bottom line: the disease itself doesn’t seem that bad in perspective (not to dismiss the potential for a lot of people to get sick, and an unfortunate number die), but the economic impacts may be tough on your 401-K, a lot of people will be hurt by an economic downturn, and there will likely be some shortages of things from China (and maybe elsewhere) due to actions (and over-reactions) by authorities and breathless reporting that is likely to kick in to high gear today.  I know it’s a lot to ask, just be sensible 🙂

Addendum: some links for information, and an analysis of the Diamond Princess data.

Do satellite images show the Chinese burning thousands of bodies outside Wuhan?

Short answer: No.

Longer answer:  No, here’s the background:  There are a number of reports floating around that are implying that satellite data is showing high Sulphur Dioxide emissions outside Wuhan, China,and that means they are secretly burning thousands of bodies.  It’s being widely reported, especially on “alternative” news sites based on images from “windy.com”.  Here one such image (screenshot from windy.com):

Well, it’s just not true.

There is a huge amount of satellite data available on-line.  It takes a lot of experience to interpret this data.  Unfortunately, there are people who think they can just download it and make pretty pictures, and understand what is going on. (There are also groups who know better, but do this to deliberately spread misinformation/propaganda  – looking at you, Bellingcat).   In this case, what the satellites are seeing are emissions from the huge steel mills located just outside the city.  Here are the hot spots from the VIIRS satellite for yesterday (February  9, 2020), overlaid over an October 17 2019 satellite image from Google Earth (NOTE: The image is static; the satellite hot spot overlay is not .. see addendum below).  You can clearly see this in a big industrial complex.

And for comparison, same background, but using Feb. 2017 (three years ago).  Virtually identical hot spots … 

Sulfur Dioxide emissions are in fact up, but that is because the kilns have to be kept hot, and emissions during a production shutdown are different from that during operations.

The Chinese are not exactly being honest and forthright about what is going on with respect to 2019-nCOV (the “coronavirus” outbreak there).  In the past they have often downplayed the impact of natural disasters, especially casualties and economic impacts.  That lack of transparency leads to crazy rumors, and this is yet another example.

 

Addendum:  The hot spots do change from day to day depending on plant operations.  Here is what it looked like last year … only couple of hot spots, but both in consistent locations with 2017 and 2020 …