Friday, March 27, 2020

Surviving the Coronavirus: Part 7

How Bad Could It Be? Edition

In Part 3 of this august series, the part that I have recently labeled as the Diamond Princess Edition, I respond to the claim of a Stanford epidemiologist that perhaps, maybe, kinda, the coronavirus is not so bad after all. In summary of that edition, the Stanford epidemiologist made a needlessly crappy assumption, ended up with a needlessly crappy result, and needlessly crappily misinformed his readership. I've since seen his needlessly crappy result quoted by others hoping that they can hope this problem away.

Today I see some similarly crappy analysis by John Lee, writing for The Spectator: "How Deadly is the Coronavirus? It's Still Far from Clear." John Lee doesn't claim to be a Stanford epidemiologist. Instead he claims to be "a recently retired pathologist and a former NHS [Britain's National Health Service] consultant pathologist." I guess I'll be punching up, again.

In my apparently-not-so-humble opinion, the former NHS consultant pathologist goes off the rails early in his article when he writes:
The simplest way to judge whether we have an exceptionally lethal disease is to look at the death rates. Are more people dying than we would expect to die anyway in a given week or month?
Apparently, woefully, the recently retired pathologist failed to read Part 1 of this august series, the part that I have recently labeled as the Deadly Peril Edition. Had he, he would have understood that the death rate, more properly the case fatality fate (CFR), is only one of two critically important pieces of information we need to make a first order assessment of how bad a pandemic might be. The recently retired pathologist made the same stupid mistake as did the Stanford epidemiologist. Both considered the CFR; neither reasonably quantified the eagerness with which the Wuhan coronavirus spreads.

Because I have higher regard for the intelligence and moxie of my readers than the former NHS pathologist apparently has for his, I'm not going to sugarcoat it. One can make a first order assessment of how bad a pandemic might be once one knows two characteristics of the bug: the CFR and Ro, also known as R zero or R naught, also known as the basic reproduction number, also known (by me) as the starting reproduction number.

The reproduction number tells us how many uninfected people an infected person will infect. The reproduction number varies over time depending on the number of people who are immunized (naturally or artificially) and on social behavior response. While the reproduction number is greater than 1, the number of infections will grow. When the reproduction equals 1, the number of infections will remain constant, as new cases exactly equal the number of cases resulting in cure or death. While the reproduction number is less than one but greater than zero, then the number of infections will fall. When the reproduction number equals zero, the bug can no longer spread.

If the bug can no longer spread because society has reached herd immunity (naturally and/or artificially) then the reproduction number will be permanently equal to zero. If the bug is no longer spreading because people are hiding in nooks and crannies (as they should until a decent treatment is available), then the reproduction number will be temporarily equal to zero. The bug is still out there, lurking, waiting for people to re-emerge, waiting for winter weather, waiting for people to stop wearing masks and stop washing hands. As long as the bug is out there, and as long as we don't have natural and/or artificial herd immunity, the reproduction number will increase, flatten, and decrease in waves as environmental and social conditions allow.

In terms of assessing the potential mortality of a virus or a bacteria, we are interested in a specific reproduction number. We are interested in Ro, the starting reproduction number, the value as it exists early on, when people don't recognize the disease as a pandemic in the making, before people begin to scrub up, mask up, hide in the nooks and crannies, before they become immunized, naturally or artificially. We can estimate Ro for an specific disease based on the experience of a specific population. Based on the unfortunate specific population of the Diamond Princess floating petri dish, the estimated Ro of the Wuhan coronavirus is 2.28. (See Diamond Princess Edition.) Based on the unfortunate specific population of Wuhan, where the coronavirus first struck with fury, the Ro is 2.28.

Once we have the starting reproduction number, we can determine the percentage of the population that will have to become immunized (either by surviving the infection or getting an effective vaccine) in order to created herd immunity and permanently drive the final reproduction number to zero. As I explained in the Deadly Peril Edition, the calculation is simple.
Herd Immunity Fraction = 1 - 1/Ro
Since I would somewhat rather be labeled a shipist than a racist, I'll use the Diamond Princess Ro for the calculation. If the Ro is indeed 2.28, then 56% of us must survive an infection or be effectively vaccinated so that we can all benefit from herd immunity.

Now that we have the starting reproduction number, which gives us the herd immunity fraction, can we begin to make sense of the case fatality rate. I'll use a CFR of 0.77%, based on the South Korea experience. (See Deadly Peril Edition.)

Assuming no vaccine, then 56% of a given population must get infected in order to create herd immunity, and 0.77% of those infected will die. To express this in the form of an equation (just this once, I can stop whenever I want, I swear) I'll introduce a new term that I will call the Potential Death Toll Fraction, to which I will assign the acronym PDTF.
PDTF=Herd Immunity Fraction x CFR = 0.56 x 0.0077 = 0.0043 = 43%
"Hello. I'm The Skeptical Juror and I'm an engineer. I haven't formulated an equation since my last post."

Without identifying the loopholes in the coronavirus, and without leaping through them, then the PDTF for any given population is 43%. Given that the population of the good ol' USA is around 325 million, the PDT is 1.4 million fellow Americans dead of the virus. Given that the population of the good ol' Earth (my favorite planet) is 7.5 billion, then the PDT is 32.2 million fellow Earthlings dead of the virus.

That's nothing to sneeze at, even though sneezing is not one of the symptoms of the disease.

Here's the good news. The PDT (potential death toll) need not equal the ADT (actual death toll), and it certainly won't. To fight back, and to save more than a million American lives (and tens of millions of our fellow Earthlings), we need to:

1. Flatten the hell out of the curve. Those of us most likely to succumb to the disease and those of us most likely to spread it to others must head for, and remain in, our nooks and crannies.

2. Make available an effective treatment / cure. There are some candidate treatments out there. We need to test the most likely of them quickly. If they fail, we need to be ready with more to quickly test and distribute.

3. Develop an effective vaccine. All the smart people say this is a year to 18 months off. We can't wait that long. That's why we need the effective treatment quickly, conventional protocols be damned, full speed ahead.

If someone sees a better path forward, please let me know. Do not, however, be like the Stanford epidemiologist or the former NHS pathologist. Do not talk to me about just the death rate. I'm sick and tired of hearing about just the death rate. Talk to me about both the CFR and the R, and how we can reduce both of those deadly numbers.