Saturday, March 28, 2020

Why Isn't COVID-19 Just Like the Flu?

What's the big deal?  COVID-19?  Most people don't know anyone who is affected yet.  People die from the flu in far greater numbers, right?  This year from the flu, CDC expects about 50 million Americans to get infected, 25 million to seek medical care, 500k to enter the hospital and perhaps 50k to die.  These are huge numbers!  And yet we don't shut down businesses, schools or transportation systems.  Do we?  Millions of people aren't laid off, are we?  What's the fuss?

What IS the difference?  And why can't we simply apply the same policies and procedures we have with flu to this new viral infection?  Cities and countries are reacting when there are only tens of people dying.  Compared to 50,000 who die from flu, what's the big deal?

Well, we're arguably three full months into the crisis and a lot of Americans still don't believe it's a real problem.  And what's worse is that most of those who DO think it's a problem, can't explain it to those who don't.

Why isn's COVID-19 just like the flu?  Let's break it down and see if we can figure it out.  Here it is.  With simple numbers.   I'll try to use numbers that can be verified.

First the FLU.  I can sum it up like this in round numbers (so the relationships are more apparent):
  • If 0.1% of 50M flu infections are fatal, that would produce about 50k deaths.  
  • That assumes an infection rate of about 15% of the US population.  
  • These numbers are approximations from the CDC website.
  • Notice that CDC predicts that only 15% will be infected because the other 85% are either vaccinated, immune from past flu infections or protected by those who are protected by herd immunity.
  • Hospitals handle this load every year.
  • So distancing measures are not needed.
  • So schools and businesses remain open.  
  • No problem.
Now let's consider COVID by comparison.  The most important numbers to understand the differences between these viruses is how many people are infected will seek medical care, how many will need assistance, how many will require critical care, and then how many will die.  There is still a large range and a lot of room for error.  Estimates in Italy suggest that 8% of all COVID cases have resulted in death.  But we really don't know infection rates and how many individuals are infected without symptoms.  Low estimates from other countries which have done more widespread testing, put this number closer to 1%, still 10 times as lethal as flu.  For our purposes, let's take a lower estimate, consistent with what I've been reading about cases in NYC ICU's today:  about 2.5%

To summarize:
  • Assuming a 2.5% mortality rate, only 2M COVID cases would lead to 50k deaths, the same number CDC estimates from the seasonal flu this year.
  • In other words, only 4% of the flu cases will produce the same number of COVID deaths.
  • My source for these numbers is a podcast TWIV, "This Week in Virology."  It's awesome.  Check it out.
  • As terrible as these deaths would be, we've not actually seen the larger problem yet.
  • First, those infected who need care, even if they survive, they will put a huge load all at once on the healthcare system, already burdened with the seasonal flu.
  • If overloaded, healthcare quality will suffer and mortality will increase.
  • But even more importantly COVID is very likely to spread to more than 2M Americans if we treat it just like the flu.  (Remember, I got the 2M to produce the same number of flu deaths.)
  • 2M cases would represent an infection rate of about 0.5% of the US population.
  • BUT after taking a toll that comparable to that of the flu, THERE WOULD STILL BE 99.5% of the population who can STILL BE INFECTED... because there are no vaccines and we have no immunity.  
  • We have a problem, a problem that is a different order of magnitude than the flu.
So right away, it should be clear:  this is not the flu.  If you don't see the difference read it again.  Infecting only 0.5% of the population with COVID produces the same number of deaths as 20% of the population being infected with flu.  And the remaining 80% of the population not infected with flu are protected... while the other 99.5% of the US population is not protected from COVID.

It's not the same.

This is just fractions, ratios, percentages.  Simple, right?  Why don't we get it?  Well, in my opinion, nobody in the media is telling precisely this story in this way.  I dare you to find a single, simple comparison like this in the press.  Most coverage does not dwell on the dynamics of the problem itself, instead rushing to conclusions about solutions, and in some cases, assigning blame.  Meanwhile, there are many Americans listening to media and leaders who continue to suggest, without evidence, that COVID is no different than flu, that is "no big deal" or worse, "a hoax."  That’s why we don’t get it.

But this is why we should be concerned:  what if we cannot slow the infection down and even a small percentage of the remaining 99.5% of the population were to get infected?  The numbers get really big, really fast.  Hospitals are overwhelmed.  Care suffers.  Doctors and nurses die.  And the situation deteriorates.

This is not theory.  We have evidence:  China, Italy, Hong Kong, Singapore and S, Korea all demonstrated the same dynamics of the disease itself, although they were treating it differently.  In Wuhan, for example, only about 0.11% of their population was infected before their distancing and lockdowns started to produce results, and it nearly destroyed their healthcare system.  This is why:  it all comes at once… and lots more people get really sick for many weeks.  If you don't care for them, they would die too.  In other words, time and resources matter too, not just ratios or percentages.

Now, lets talk about ratios, time and resources — the SAME ratios but different numbers.  Keeping round numbers, here is another scenario focused on today, March 28th, 2020, and the immediate future, one we CANNOT CHANGE.
  • NYC hospitals are completely saturated right now, 100% occupied as of yesterday.   
  • In some hospitals, there are already two patients on a ventilator. 
  • In 10 days, all these numbers will double two or three times.  (Remember, the rate of doubling is 3-4 days and there is a three week delay for infected individuals to seek healthcare.)
  • In only 10 days, NYC will need 4 to 8 times the total number of beds, staff and supplies than they have right now.  
  • Based on this rate of doubling, it’s reasonable to assume that there will be about 500k people infected nationwide with COVID in next 10 days.   But we cannot see them.  And they won't show up until April 11th.
  • 20% of them or about 100k people be receiving or seeking healthcare, up to 8 times local capacity, depending on where they are on the curve right now.
  • About 5% of all cases will need intensive care, or 25k people for several weeks of intensive care, up to 8 times capacity today.
  • Many of them will need ventilators, up to 16 times capacity today (remember, ventilators are ALREADY oversubscribed).  
  • And half of them, 12,500 will die, many of them healthcare workers.
  • My source for these numbers is a podcast TWIV, "This Week in Virology."  It's awesome.  Check it out.
Now let's look just a little bit further out and think about doing NOTHING.  And if we do NOTHING, consider how our actions today might actually make a difference in the coming weeks.
  • Based on how many people are hospitalized today and who have died and the advice of experts, assume that there are approximately 100,000 people infected in the US to consider our choices.
  • If we do nothing, this number is doubling every 3 or 4 days.
  • This will continue until distancing and lockdowns start to work, perhaps 2 or 3 weeks from now (because there is a delay between infection, the appearance of symptoms and seeking help).
  • Consult TWIV.  I'm telling you, it will not disappoint.
That means that -- if this assumption is correct -- without distancing measures in place, in only 20 days it would double 5 times;  we’ll have about 3.5 million people infected or 0.1% of the total US population.  But in 24 days -- only 4 days later -- it will be 7 million people and 0.2% of the population, twice as many admissions, twice as many ICU patients, twice as many fatalities.  You get the idea.

Read that paragraph again.  Seriously.  Read it again.  It's not "like" the flu.  We'd better treat it differently.

Notice that I said IF.  It is an assumption.  We can change that number if we act RIGHT NOW.   (And by 'we' I mean every single American).  We could change that number were we to start distancing with a stay-at-home advisories and more aggressive lockdowns right now.  We'd need the support of the Federal Government and every State Government, most likely.  The support of school board and business leaders will be important because we need to close schools, non-essential workplaces, and cancel large gatherings,  We need country-wide, clear instructions about not traveling and not working if you're sick with provisions for sick pay, free testing and free treatment.  (Otherwise it doesn’t work;  people don't stay home).  Congress has already delivered a lot of this.  (Could have been a week or two earlier, but I'm not complaining).  Now we just need the White House and a few more governors...

Lots of people are worried about models.  What is a model, actually?  What assumptions are behind the model?  Is it useful if the assumptions I've made a wrong?  Can we trust it?

We should trust this simple model -- a number model based on simple arithmetic -- because its purpose is very simple and we're transparent about that purpose:  it's designed only to show us that COVID is not the same as flu and therefore we should be receptive to interventions such as quarantine, isolation and social distancing that we do not deploy with flu.  The purpose of the model is to help us understand the relationship between the spread of COVID and limitations on our healthcare system.  And it's pretty easy to be confident that if we do NOTHING, we'll be overwhelmed.

The number models with ratios and time I used above do not predict what people will do.  They only show us what the virus does, assuming we treat it like the flu.  Nor do they tell us what we should do in response to the virus.  There are no assumptions at all about solutions to problems, policies in these simple models. The model is not a crystal ball.  It is a tool to understand our choices and the implications of our choices.  We do not have to tread COVID like the flu.  

The only policy changes we can make in the short run will CHANGE the rate of infection and the AVAILABILITY OF RESOURCES we can affect immediately.  Entire new hospitals are coming on line and thousands of skilled people are mobilizing to help out.  Resources are late... but they are coming, thanks to thousands of people who have known all of this for weeks now and, despite contradictory messages from our leaders, are producing facilities, staff, beds, protective equipment and ventilators.

The infection rate can also be affected with policy changes now.  Although we cannot see the effect yet (because of delays in the appearance of symptoms and seeking healthcare) it is has probably started to slow as of last week when people started getting serious about distancing.  Perhaps two or three weeks from now, AFTER we see the effect of distancing or a total shutdown, well see the rate of increase start to level off…. AND THEN we’ll need another 4-6 weeks (or more, I'm not too clear on this) to treat those people until the resources in the hospitals can be reallocated.  This is what happened in China.  The entire process took a couple of months to develop and be resolved.  That's how long it will take here too, IF WE ACT NOW.

If we delay it will take even longer.

If we relax our policies on distancing too early, it will reemerge and it will also take even longer.  Remember that if we return to everyday activity while the virus is still active in the population, it will start to double in the still unaffected population, which will still be the unprotected 99.5% of the US population.

One of the big uncertainties is the rate of infection.  Until we can do broad-based testing of the population.  What if we're off by a factor of 10?  What if 5% have been infected and we're seeing only the symptomatic individuals?  It means that 95% are still unprotected.  It makes no difference to what we should actually do:  isolate ourselves.  That's what the model is for, to ask the question, "What if?"

Easter?  I don’t think so.  But let’s play it by the numbers, shall we?

This is not the flu.  Once the flu affects 20% of the population, it's DONE.  This is COVID.  Now we know the difference.  Once we allow this wave to subside, hopefully preserving our healthcare infrastructure, will we still be vulnerable to another wave, until we have a vaccine.

I don’t know why this is so hard to get, on the one hand.  But I've honestly never seen it presented with the ratios laid out, with time and resource dimensions added and the fact that 100% of our population is vulnerable, unlike with the flu.  And it's 10 or 20 times more lethal.

There are only four things that matter, and we can see and measure them ALL:

  • hospitalizations;
  • deaths;
  • rate of doubling;
  • limits in the healthcare system.

We can SEE IT.

This is not a planning problem, something we didn’t know about or prepare for.  It drives me crazy when I hear, "How were we supposed to know?"  There are thousands of experts at CDC and hospitals around the country who knew absolutely, without a doubt, by the end of January, that if we were unable to initiate distancing measures consistently until mid-March, this is EXACTLY WHAT WOULD HAPPEN.  These experts are watching a slow-motion train wreck they’ve seen many times in their minds.

Truth is, as a society we don’t recognize these experts.  We ask them to have a plan.  They planned.  We asked them to advise us.  But then we don’t listen to them.  We don’t trust them.  And we don’t allocate the resources we should to address public problems like this.

This is an execution problem:  we didn’t execute the plan.  We didn’t listen.  We didn’t allocate the resources.

BTW, there are other problems EXACTLY LIKE THIS.  Climate change anyone?  It’s a train wreck emerging over decades instead of weeks.  But it’s the whole freakin’ planet.

No comments: