One curve to rule them all (and why it doesn’t exist)
8 weeks ago the world was paralysed, petrified by the possibility of pandemic catastrophe, where the entire world looked like Wuhan writ large. We saw the images coming out of Bergamo and Queens, and much of the world mobilised to make sure other societies escaped the same fate.
The experiences of these past months have left us with as many new questions as they’ve given us new answers. Why has Sweden managed to get away with a relatively open and unchanged society, whereas the UK’s brief experiment with that model was a swift and utter failure? How did some parts of the US experience near armageddon, while others have had minimal distancing policies but also minimal disease?
What we know today, is that different societies have very different levels of susceptibility to Covid-19, causing them to behave in very different ways. There is no single “natural history” of Covid-19 in a community. The way a disease spreads varies enormously based on local conditions, and our response to the disease will have to be equally varied — over time, and between places.
Think about the susceptibility of a community like the responsiveness of a gas pedal in a car. Even a light tap on the pedal in a Ferrari will send you rocketing down the highway, while you have stomp on a dump truck’s accelerator for 5 minutes just to get out of the parking lot. It is the same for communities with Covid-19 — highly susceptible communities will incubate and amplify a small number of cases extremely quickly, while low susceptibility communities will have a lot more time to react, and not need to brake (by applying shutdowns and social distancing) nearly as hard.
The most important factors which determine a society’s susceptibility seem to be:
- Population density (high is bad)
- Inter-generational connectivity (more elderly people living with and interacting with young people is bad)
- The general age and health of a population (more elderly, with more comorbidities is bad)
- Connectivity with outside communities (higher is bad)
There are likely others, based on weather conditions, air pollution, genetics, and other factors, but we don’t know enough to say for certain which ones are significant, and how to weigh them.
This differentiation plays out even on a micro scale — even as Queens has been overrun with Covid patients, next-door Manhattan is much less affected. In Massachusetts, there is a 10x difference between high incidence and low-incidence towns, despite similar lockdown rules being applied to the entire state. We can see in the below graphs that population density and poverty (which is related to both inability to socially distance as well as a higher likelihood of multiple generations living together), seem to drive much of this variation.
This is part of why classic epidemiological models are proving themselves so useless — there is so much micro-level variation that looking at and predicting patterns at a state or national level becomes a meaningless abstraction. It’s only when a highly susceptible population has a big influx of infected people and does not aggressively distance (either by policy, or because people cannot afford to because of their financial situation) that we see the catastrophic waves like in Bergamo, Wuhan and Queens.
In general, societies will fall on a spectrum between two big pattern types, depending on the susceptibility of their population.
- The SeeSaw: Concentrated populations with multiple medical problems living inter-generationally will follow a saw-tooth pattern over the coming 1–2 years, as huge spikes in disease are followed by a tight lockdown. The epidemic gradually wanes to a point where it is mostly under control, society re-opens, and unless that opening is very tightly calibrated the infection resurges in this susceptible society. This pattern then continues until herd immunity forms or a vaccine is widely distributed. This is what we can likely expect in places like the outer boroughs of New York.
- The Slow Burn: Societies with few risk factors will have an undulating level of disease operating in the background, but it will rarely get out of control. With adequate testing and surveillance, approaching outbreaks are managed proactively with less severe lockdown measures, and a baseline level of Covid activity, in the setting of more moderate social distancing, becomes the new normal. This is what we can expect in environments like non-urban Sweden.
Societies in between will need a level of restriction matched to their level of susceptibility and the current level of disease transmission (which will need to be tightly monitored through testing in any strategy). This hyper-local constantly recalibrated approach will require a lot of close management, with government officials and corporate managers constantly reassessing risks and levels of disease spread to micro-target interventions. The concept of community susceptibility is sadly unaddressed in the EU, CDC and White House guidelines on “reopening” — by focusing reopening criteria only on levels of disease activity in a certain area at a certain time, and not the underlying susceptibility of particular areas, the current approaches will over-restrict some environments, and under-restrict others. California’s county-by-county approach is a notable exception, but it’s unfortunately unique.
This is going to be exhausting for everyone — but neither indefinite “lockdown”, nor immediate “liberation” are feasible. A middle way, requiring thoughtful and constantly updated consideration of local conditions, is the only way we will get through this pandemic with our societies intact.
