The data behind the case for reopening Schools and Businesses, now
The only positive from our collective failure to contain the novel Coronavirus is that we now have data and can move from fear-driven decisions to data-driven decisions.
From the beginning of the epidemic, we knew that the virus affected people differently; we now have data that can help us refine our strategies. I used the CDC data to compare the causes of death for all deaths from 2/1/20 to 4/25/20:
Notes: the death rate of the flu is much lower in 2020 because the Covid containment measures also slowed the spread of the flu, so I plotted the flu death rate for a typical year (2016 is the latest year available on the CDC database), as a reference point. Also note that the death rates for accidents and the flu are NOT annual rates, it is the average death rate over a 13-week period (based on the 2016 annual death rate)
A few basic observations:
- The death rate of Covid for people below 55 is similar to the one of the Flu and overall very low. People below 45 are 14 times more likely to die of an accident than they are of dying from Covid.
- Covid is much deadlier than the flu, period.
- For people over 65, and especially people over 75, Covid is very deadly, much more than other causes.
Those numbers are national averages; they are biased since the virus has not widely spread in most places in the US. We therefore need to look at places that were hit harder (such as New York City or Italy) to get additional insights on the impacts of the virus as it spreads:
From that data, we can understand that:
- The virus is very contagious and can spread fast in densely populated areas with limited containment measures. The latest data estimates that at least 20% of the NYC’s population has now been infected.
- The death rate gets much higher as the virus spreads (NYC’s death rate is 15x higher than the national average). However, the death toll continues to fall almost exclusively on the elderly. In Italy, more than 95% of the people who died from Covid were over 60, and the median age of people dying from Covid was 81 years old . In NYC, the death rate for people above 75 years old is 74 times higher than for people 18 to 44.
- While the death rate for people below 45 years old remains much lower than for older people, it is no longer negligible in NYC (15.1 deaths per 100,000). That Covid death rate is comparable to the death rate of accidents (15.1 vs 12.6).
None of the data here is normalized for underlying medical conditions, which are a known risk factor.
Looking at the data holistically, it seems obvious that a blanket stay-at-home strategy is sub-optimal. There are clearly 2 categories of people: older people and people with pre-existing conditions; and the rest. The first category is extremely at risk and must be protected a lot more than they currently are. It is highly inefficient to apply the same measures to the entire population when one category is 74 times more at risk.
Put differently: if you could invest in 2 different assets, one with a return guaranteed to be 74 times higher than the other, wouldn’t you invest more/most of your money in the high-return asset?
For people below 45 and/or without pre-existing conditions, the death rate is low enough that individuals should be able to decide how to manage that risk. Drawing a comparison with car accidents: we rely on individuals and society to apply common sense measures to mitigate the risk but we are not outlawing cars or trying to prevent all car accidents. We should do the same for Covid:
- For schools: Children are entirely spared by this virus and most parents of school-age children are low risk (most of them are below 50), so why keep the schools closed? School closures are a major disruption and seem completely unnecessary when looking at the data. Schools should reopen now; parents and teachers can decide if they feel comfortable going back to school (I definitely would).
- Similarly, most businesses should be allowed to reopen (assuming they follow common sense prevention measures) and individuals can decide how to protect themselves based on their particular situation.
To be clear, I am not advocating for the herd immunity theory or the end of all containment measures, I am saying that we should focus all our resources on the at-risk population rather than apply blanket measures to everyone, we need to give people data to help them understand their risk factors and let them apply the measures that they feel comfortable with.