The roll of the COVID-19 dice

      3 Comments on The roll of the COVID-19 dice

“It doesn’t want to kill you before you transmit it.” – biologist Michael Farzan of Scripps Research

That quote is not so scary as it sounds; it is basic probability math. The probability math of the COVID-19 Coronavirus variant has hit the news, and not in a good way. We have already seen the first reported cases in my county in Florida. Here are a few thoughts on how to rationally evaluate the risks you face with this new virus.

Two basic risks – virulence and death

Some medical threats, whether bacterial or viral, do not spread easily. For instance, the HIV virus, contrary to initial fears, proved not particularly “virulent” outside of a few high-risk transmission “vectors,” particularly unprotected sex, blood transfusions, and infected needles. And so, most of us have been, probabilistically-speaking, “lucky” to avoid that threat.

The virulence of COVID-19 is perhaps the biggest unknown at this point (and thus the reason to stay tuned for the “Be a Bayesian” discussion below). How long can it be contagious before symptoms appear? How close do you need to be to an infected person? Are some people more susceptible to infection? These are the questions for which you need to seek from reliable news sources. The more “reality” you know, the safer you are. And bad information can put your health at risk.

“Community spread” occurs when a virus replicates easily and sustainably through “normal contact” with other people, mostly simple touch and “breathing distance.” The Centers for Disease Control have now rated COVID-19 a “community spread” virus. A statistic called the “basic reproduction number,” also called R0 or “R-naught” measures the average number of people you might infect. Measles has an extremely high R0 of 18, while the seasonal flu’s R0 is usually around 1.3. The latter number, for instance, means that any 10 infected people are likely to infect 13 more people. [1]

Watch this number! The World Health Organization has pegged the R0 range for COVID-19 as between 1.4 and 2.5, but that number will change as more becomes known. R-naughts above 1.0 can quickly turn the spread of a virus exponential, and so far, this appears to be what is happening with COVID-19. Because of that exponential nature, seemingly small differences in this number have dramatic effects on the growth in the number of infected people. At some point the exponential growth peters out, but the U.S. appears to be a long way away from that slowdown if this infection follows the course of China or Italy.

The case fatality rate (CFR) is better known, although still in flux, based on how many people present themselves to medical facilities with advanced symptoms and how many of those people subsequently die. The current odds here appear to be 2.3%, or about 2 out of 100 infected people. Depending on your perspective you might perceive number as low or high. It’s high. From one nursing home in Kirkland, Washington, at least seven people have already died.

In comparison, the seasonal flu annually kills at about 1/20th that rate, or 0.1%. In both cases, there are factors that put some people at much higher risk than others, for instance advanced age, compromised immune systems and other poor health factors. While you may not be in one of those risk categories, you likely know someone who is, and so you do have a moral responsibility to avoid being a carrier if at all possible.

And so, the overall risk is some combination of these two or more “dice throws.” In fact, they influence one another. If the case fatality rate seems high in Washington State, it is likely because we have underestimated the R-naught there. A major reason more cases have been found in Italy and South Korea is because they have been vigorously testing for the virus. And in the US. we are not. Time will change the odds and make them clearer, but ignorance here is not bliss. [2]

Be a Bayesian

Bayes Theorem

I have written in the past on the concept of Bayesian inference. This is a time-tested process for getting a handle on risks in situations when you don’t know the exact risk facing you, but when you are constantly getting better information. I won’t do the math here, but the idea is “the odds” in an unknown situation are really a quantification of your confidence level in the accuracy of the information. As noted above, this is the current state of our understanding of Covid-19’s virulence. We just don’t know a lot yet.

Bayes Theorem is an “iterative” method where you continuously refine your initial guess (the “prior”) with new information as it comes in, producing an updated, more accurate guess (the “posterior probability”). Those updated odds then get looped back around to be mathematically combined with any new data as it comes in. This is basically how weather forecasts get more accurate, both specifically as a storm approaches and more globally as new storm seasons come around annually.

When you “don’t know what you don’t know,” you have what is called an “uninformed prior,” and the usual response is to plug in 50-50 odds in these cases in order to get the loop started. Let me suggest that, the less you know about COVID-19 the more the prudent position is to place your personal bets at the 50-50 level. I don’t know about you, but when bad things might happen at the odds of a coin flip, I personally get much more cautious and act to honor that caution.

I wish I could say that I had confidence in the information coming from the top of the current administration, but these guys are the epitome of “uninformed priors” much of the time. The Centers for Disease Control is suffering in real time from budget cuts and preparedness mistakes, but at least there are respected scientists still working there. The World Health Organization has also been a reliable source for accurate news on this topic. “Be a Bayesian” and keep informed. But recognize that some news channels may make you more ignorant.

That quote

That opening quote, “It doesn’t want to kill you before you transmit it,” requires a bit of clarification. Viruses do not have intention; they are not out to kill you. They are out to survive. Or more accurately, any virus mutation that does not find ways to reproduce itself will quickly disappear. This is the same basic evolutionary math that if your parents did not have any children, then neither will you. Viruses do not have conventional “parents,” but perhaps you get the point.

What biologist Michael Farzan is expressing here is that if a virus is “too good” at killing off its host, it will not get a good chance to spread. This is what has “saved us” thus far from the Ebola virus, with an R0 of only 2. Its victims die quickly, albeit in messy ways. But if you can contain that mess, however, the virus dies before spreading.

The danger here, then, comes when a virus “hides” in you without visible symptoms long enough for you to come into physical contact with several other people. A long incubation period mathematically increases the chance that the virus will replicate in other hosts, and that is a major threat with COVID-19.

God will protect me

The Shincheonji Church of Jesus, a Korean religious organization with over 200,000 members, has been demonstrated to have been a major “vector” in the rapid spread of COVID-19 in South Korea. As of the end of February, most of the Koreans infected by the virus were members of the church. The organization’s practice of close physical contact during worship services, its cross-country proselytizing, and its secretive nature apparently allowed the virus to gain a foothold before being detected in the broader population.

Vice President Pence, who is heading up the Administration’s strike force on COVID-19, has a history of injecting his personal religious beliefs into public health issues while he was governor of Indiana. His moral discomfort with HIV-related issues reportedly delayed implementing effective preventative measures during a major HIV outbreak in his state in 2016.

I recently wrote a brief update on the concept of theodicy. This is the “Big Question” of “the justice of God” that comes into play when people of religious faith confront big life challenges that threaten themselves or their families. My best theological advice would be “Do not put the Lord your God to the test,” and instead wash your hands frequently. And keep the Hell away from me if you are sick.


Notes:

  1. The seasonal flu is “nothing to sneeze at.” Reports of deaths so far this season relating to the flu are in the range of 18,000 Americans. Like gun deaths, we have become so inured to this number that we fail to realize how easily these statistics could be changed. Get the damned flu shot!
  2. The danger here is that, the better we are at staving off the results we are seeing in China, the more that less-informed people will see this incident through a “hoax” lens.

3 thoughts on “The roll of the COVID-19 dice

  1. Lis7

    “… but these guys are the epitome of “uninformed priors”….”
    Must have been going through the article a little too quickly, because the first time I read that sentence I thought it said “uninformed prions”. I’m sure implicit bias on my part had nothing to do with that, however 😉 .

    Reply
  2. Pingback: COVID-19 and real-life lifeboat ethics – When God Plays Dice

  3. Pingback: Kids, can you say Fecundity? – When God Plays Dice

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.