In most professions, excessive caution in giving a range out outcomes can be as wrong as a lack of it. In medicine, telling a patient about a remote cancer diagnosis without adequate data could be considered negligence if the patient suffers nervous shock. In law, a lawyer who causes a real estate closing to fail based on a remote title issue could be held liable when the value of the property increases.
In recent weeks, governments in the Western World have published pandemic modelling offering alarming scenarios of the spread of the Covid-19 virus. The scenarios involve conventional medical responses with minimal measures in communities to curb the spread, in contrast with community lock-downs or “social distancing” measures. In Ontario, Canada, for example, public health officials published a slide deck with the figures informing the provincial government. It offered a grim prediction as of April 3, 2020, of 300,000 cases and 6,000 deaths by month’s end. The same modelling anticipated 80,000 confirmed cases and 1,600 deaths, even with emergency measures such as the closing of schools and non-essential businesses. There has also been criticism of public health agencies’ implementation of broader testing capacity.
Ontario is at mid-month. More testing has not really found increased rates of infection, and deaths have remained steady or have declined. We can only rely on the published figures. Even if we were to increase testing to full capacity, it is hard to imagine the active case load increasing from about 4,000 to 80,000 in two weeks, and the deaths increasing from 291 to 1,600 in the same period. The time for doubling o the epidemic in Canada has increased from about 6 days to 10.5 days (New York Times). This is important because the shedding period for the virus is about 10 days, and on average seven days or less. If the virus does not find an equivalent number of new hosts to infect within 10 days, the epidemic starts to decline. The larger the number of days in the doubling figure, the more rapid the decline. (Eg., in a human population, if couples do not have two or more children during the ordinary 25-year period of fertility, populations decline. In the case of viruses, we count generations in days.)
The likelihood of anything happening by community spread depends on a host of factors. While it is true that one can contract the virus by being in close contact with an infected person (eg. by eating together at a restaurant), and this coronavirus is more virulent than the flu or the cold, the mechanism of infection remains the transfer of viruses, either from expectorant (coughing, spitting while speaking) or from surfaces or handshakes. The virus does not transmit very well through the air, we are told. It is possible, but not very likely, that you will contract the virus by passing someone in the street. We remember from high school science that air molecules are less densely organized than liquids. Imagine a room full of balls at Ikea. A child with a cold can infect others in the room over time, by playing with the balls and by the others playing with the same balls. This will not likely happen instantaneously. Over the course of an hour of playing, many of the children will have caught the cold. Multiply the balls by a factor of ten, and it would take ten hours longer to infect the other children and the likely result is that none will go home with sniffles.
This analysis, however, crude and technically flawed, is nevertheless more accurate and reliable than the modelling produced by our finest minds with training and qualifications in epidemiology. It is hard to argue with what is actually happening. A graph comparing the worst-case, best-case and actual cases of ICU use in Ontario, shows the actual figures consistently about five times better than than the best projections of the published model. It is possible that viral generations, like the fragility of life in general, is susceptible to extinction dynamics when conditions are less ideal for transmission and regeneration. Not enough yeast in the dough will ensure the bread comes out of the oven hard as a rock.
Only a multidisciplinary team of scientists could really tell us why the models have been so wrong, but one thing is clear: not enough work has been put into the underlying assumptions. The scientists have to be honest with us: if it is like predicting the weather a month away, then tell us. The likelihood of hurricanes has not warranted permanent evacuation of Florida, but it has informed the need for shoring up disaster response capacity and keeping shutters in the shed.
Don’t tell us the figures are so much better than the best case because people are observing the rules, after you told us the best case if everyone observed the rules. When official figures show an overall decline in deaths from all causes over a similar period, one has to admit how wrong the projected effects on public health have been. Rather, admit you were wrong (very wrong) and help us get through this because we all still rely on you.
It would be unfair to impose standards of proof as we do in the law courts. We would not apply the Daubert rules to public health information intended to guide people and governments through these extraordinary times. But it is also wrong to sensationalize data and science in order to undermine the proper use and accountability of government authority. Diversion of vital health care resources to appease myopic experts will cause suffering and death to those who cannot access cancer surgeries and other vital non-Covid-19 ailments. While fear and greed are the eternal political motivators, politics must not so overwhelm and distort the truth or it will lead to distrust and cynicism when the predictions turn out to be so wide of the mark. The truth, at all times, must be the measure of reasons why we do or do not do things.