Coronavirus: uncertainty and assessment of risk
Updated: Apr 22, 2020
Fear, uncertainty, and social media are a bad combination. A perfect storm for misinformation and mass hysteria. During this outbreak of Coronavirus, I highly recommend you consult only trustworthy resources such as the CDC (Centers for Disease Control) website. They provide a situation summary, updates, what is known, and what is not known.
Most other sources suffer from conflict of interest. Although they want to keep their readers informed with accurate information, they are also driven to increase traffic to their sites. Traffic is accomplished by providing information that is novel and important. However, when it comes to this new Coronavirus, there is currently very little to report - mainly uncertainty. It is a new mutation and new mutations have different characteristics from older versions of the virus. These new characteristics need to be determined. In addition, we are at the mercy of the Chinese government for information - which is not known for its openness. Hence, it is almost impossible to make future predictions about this virus.
With a captive audience, media outlets must put out something. Their solution is to lower thresholds for what they consider acceptable evidence. They will use lower forms of evidence such as anecdote and expert opinion. It’s not that there is anything wrong with speculation, but speculation must be accompanied by full disclosure. An explanation of how the new evidence fits into the evidence hierarchy, the source of the evidence, the quality of the evidence, and how likely is it to be accurate. In reality, this is rarely done. Ideas based on weaker forms of evidence are presented on equal terms with those based on stronger forms of evidence.
To make accurate predictions for the new Coronavirus outbreak we would need the following:
A rapid, accurate, and widely available test for the disease.
Determine the natural history of the disease and case fatality rate. What are the symptoms of the disease, what percentage of cases are fatal, and who is at risk? The SARS outbreak of 2002 (another Coronavirus mutation) had a case fatality rate of about 10%. Case fatality is hard to determine initially because we generally know the number of deaths but we do not know the number of patients that have mild or asymptomatic disease.
Determine the mode of transmission. Is it spread through the air or only by direct contact? Can it live outside the body? on surfaces? and for how long?
Determine the Incubation period. What is the time from when the virus is contracted to when the patient exhibits symptoms?
Determine the period of viral shedding. At what point does the infected patient become contagious and at what point does she cease to be contagious?
There are a few general principles in infectious disease. For instance, the more lethal a virus, the less likely the virus is to become widespread. From an evolutionary perspective, it is not advantageous for a virus to kill its host as it limits its ability to spread. Ebola virus is a great example - highly lethal (63% case fatality rate) but containable. Yet, by and large, most of the characteristics of a new virus need to be calculated by the retrospective and systematic analysis of prior cases - a process called descriptive epidemiology. There is little-to-no human experimentation. For instance, the simplest and most direct way to determine whether the virus can spread through cough droplets is to have healthy subjects breathe in the air in the presence of a sick patient. This kind of experimentation is considered unethical for obvious reasons. Thus, we can only guess whether it can be spread through cough droplets by the retrospective and careful analysis of hundreds of cases. A laborious and imprecise process.
One thing is for certain, we humans are very bad at evaluating risk. We concern ourselves with new and dramatic things far more than common and mundane things. For instance, we worry about terrorist and shark attacks far more than we worry about driving our car on the freeway; yet, we are much more likely to die from a car accident - maybe by 5 or 6 orders of magnitude. We worry about the new Coronavirus far more than the flu, despite the fact that there have been at least 22 million cases of the flu, 210,000 hospitalizations and 12,000 deaths so far this season. My question to you: if you are so concerned about a lethal viral pandemic, did you get the flu shot this year?