Take the red pill
There’s a lot of debate on the “real” numbers of the COVID-19 pandemic. The “real” fatality rate, the “real” number of infected, the “real” timeline of when it first arrived in the USA. Most often, we are being told the real in a way that supports a narrative of, “it’s not really that bad”, or “the worst is likely behind us”. Let’s break some of this down.
The Real Fatality Rate
The fatality rate is simple to calculate: the number of deaths (numerator) divided by the number of confirmed cases (denominator) expressed as a percent. It’s also important to note that once the “real” fatality rate is determined, it should remain fairly fixed and should not be changing significantly. This is a rate, not a count, and it then may be used as input into various prediction models. In simple terms, if the rate is 1%, then it means we can expect for every 100 cases, 1 person will die on average.
From the very beginning we’ve been told the real fatality rate may be much lower than current data is showing, and the fatality rate early on was so high because of limited testing. In other words, the denominator is not “real”, it’s too low because we haven’t tested enough and we don’t really know all the “real” cases out there. Since then, we’ve tested a lot more and have counted significantly more cases than in March, and even early April. Now the narrative has changed a bit more, and we are hearing a lot about the unknown cases, especially of the asymptomatic carriers. We need to find them and count them, too, and if we do that the fatality rate will be even lower, some have conjectured it could even be as low as seasonal flu. Mathematically, this is true. If we can make that denominator larger and larger, the fatality rate will get lower and lower.
Of course the inverse is also true, and will help drive the fatality rate lower, i.e. if we can get the number of deaths to be lower, the fatality rate of course would be lower. And if we could do both at the same time, higher numbers infected and lower numbers of deaths, we could get that fatality rate way down. The easiest way to lower the number of deaths is to simply underreport them, and we’ve already had evidence of China doing just that, and possibly several other countries. Another way that number may be lower is if deaths are attributed to the primary cause, such as pneumonia, cardiac arrest, organ failure, and not the root cause, COVID-19, either because it’s just not being reported accurately, or the person was never tested and confirmed to have COVID-19. There is a flip-side to this, too, where some would argue we are just overcounting the deaths, usually because some areas are counting even suspected COVID-19 as cause of death when it hasn’t been or cannot be confirmed. This has been happening in nursing homes, and may be happening in other areas.
Now back to calculating fatality rate. To get to the real fatality rate, it’s now clear that we also need the real number of deaths and the real number of confirmed cases. But we don’t really have either of those. Here’s the most important thing: we rarely or ever have those in cases of wide spread epidemics or pandemics. That’s why the Spanish Flu of 1918 has fatality estimates of 50 to 100 million people worldwide. That’s an incredibly wide margin of error. Does anyone think we have real numbers on seasonal flu? Guess what: we don’t. Even with seasonal flu, it is estimated that as many as 50% may be asymptomatic. These are estimates because it is difficult, if not impossible, to really know.
What we can do is use the actual numbers reported and look at trends. Here are the worldwide totals, as reported to Johns Hopkins from March 9 through April 23:
In the past 46 days, the fatality rate has doubled.
In the past 46 days, the fatality rate has increased every single day except one.
In the past 46 days, the fatality rate has done the opposite of what most predicted.
Let’s look at what happens if we model different proposed fatality rates.
To get us back in line with earlier proposed estimates of 2–3% means our actual number of cases must be 6.3 to 9.5 million people, or off by 235% to 350%. Sounds reasonable.
For COVID-19 to be in line with seasonal flu, we must have upwards of 190 million cases, and cases are off by over 7000%. It’s possible.
In either case, there is a rub. To get the fatality rate down means this part of the equation, the denominator, the number of cases has to be higher, much, much higher. For that to be true, that must mean COVID-19 is considerably more contagious than first estimated (known as the R0 or R nought number). This means a lot more people had it, have it, or will have it.
The problem is that just about anything is possible, since we simply don’t know. All explanations are equally rational, but equally irrational. Dr. Anthony Fauci has numerous times reminded us that the models are only as good as the “underlying assumptions”, and he stresses that as better data becomes available the models can be updated and thus get more accurate. This is why the experts and informed leaders have been demanding more testing. We need more accurate data.
The red pill
We are at a critical junction, as many states are soon beginning to open more businesses and ease restrictions, and most states are planning to in the fairly short term. We could all take the blue pill, and wake up in the morning and pretend none of this is happening. Let’s not go that far. Let’s just wake up in the morning and pretend the best possible imaginable scenario: COVID-19 ends up being, almost, like seasonal flu and has a fatality rate of 0.1%. Well, we’re not quite out of the woods, because to get to that magic number means COVID-19 is far more contagious than seasonal flu (and far more contagious than initially estimated). Let’s say 300 million Americans contract the virus. I’ll be even more generous, and we can say 200 million get COVID-19. That means with a fatality rate of 0.1%, around 200,000 people would die. This is about 5 times more than a typical flu season, and almost 17 times more than the number who died of Swine Flu in the 2009 pandemic. How about we model 100 million Americans get sick, but the fatality rate is just marginally higher at 0.5% (still way, way lower than the actual 7%). In this scenario, 500,000 people would die. Now you can see why Dr .Fauci says the models are only as good as the assumptions.
Or we can take the red pill. We can wake up in the morning and accept the real we have to work with, and act prudently until we have more data, more facts, and a much better understanding of COVID-19 and what may or may not happen in the coming months. I am just as hopeful for good news as anyone else, but I just need it to be real. Ask anyone who’s lost someone during this pandemic, and they can show you what real loss, real grief looks like. So until real answers, real treatments, and a real vaccine are available, we must do our best to avoid the real outcomes of this very real virus.
