March 30, 2020
Decision making under radical uncertainty
We live in times of radical uncertainty. In the absence of tests a runny nose could be construed as a mild Covid-19 symptom. Others may already have had the virus but still fret about getting it. Only those with clear symptoms have some form of certainty. Now imagine the same at macro level. Classic cost-benefit analyses are of limited use as there are more unknowns than knowns.
We know that reported cases are a useless statistic. We do not even know whether people die due to Covid-19, or just with Covid-19. We may count those that would have died anyway and classify them now as Covid-19 deaths. Why is the mortality rate higher in some regions than others? A 10% mortality rate in Italy versus 0.8% in Germany raises more questions than it answers. Did the Germans get it right due to more testing? If infections are 10 or 20 times larger as some studies suggest, the mortality rate would be much lower even if we were to assume it was caused Covid-19.
What we do know though is that in hotspots our health systems will be challenged to cope with the number of patients in need of respiratory help within a very short period of time. The main argument for lockdown measures is to help health care systems cope. But how long does the lockdown have to be in place? This depends on the rate of growth of serious cases. Without testing, those numbers will be based on assumptions that may or may not be right. But can a government take the risk to find out? In times where politics is flying blind, edging towards the cautious side may be the only guiding principle at hand.
Over 60 countries in the world have now ordered a lockdown in their fight against the pandemic. It will help the health systems cope and prevents governments from being blamed later. Those who die from the negative effects of the lockdown will remain largely unaccounted for. Remember the suicide rate that soared in Greece after the austerity programme was imposed by creditors? There is a short-term bias for politicians, while the long-term costs are largely unaccounted for.
What about Japan, Switzerland and Sweden? These countries defy the general trend and also have seen their infection curve flattening. There has been no reported shortage of intensive-care beds throughout the disease-spreading period in Japan, one of the first countries to report Covid-19 cases. In Sweden children still go to school and measures are extremely light. We will only know much later whether their way of responding to the crisis succeeds right to the end of the curve. Will Vladimir Putin and Donald Trump get credit for not giving in to fear, or will they be seen as recklessly risking their people's lives?
The lockdown has costs for society. Some people fall through the net, like the self-employed in the UK who cannot expect to see any money for months. Strict confinement is testing families and relationships to their limits. The fear of contagion might also compromise our ability to invent and persevere in times of adversary conditions. Health issues further down the line are likely. Those costs are carried by a population that exceeds by far those considered to be at risk from the Covid-19 virus. Trade-offs will have to be made, especially as the lockdown continues for months. And this is likely to be painful either way.
The only way to bring clarity is through testing. But the critical issue here is time. Developing accurate, reliable, validated tests is difficult and takes time, warns John Lee, a pathology professor in his analysis for the Spectator. At the moment the public is taking the tests at face value, trusting that they are measuring what we think they are measuring. But this might turn out to be wrong, as tests might not distinguish clearly between Covid-19 and other viruses, and may not be sensitive enough to pick up all those infected whether they display severe or only mild symptoms. There is also a lack of tests. Many countries stopped testing or limited them to health care workers. This means that many of those tested positive already represent a biased selection of the population. We also will not know for some time whether people had the virus or not. Antibody tests are unlikely to be rolled out any time before May.