Preparing for the storm – gathering information about COVID-19

Jonathan Dushoff, 2020 Mar 23

COVID-19 presents humanity with a new kind of challenge. We’re faced with a pandemic that could kill tens of millions of people. But we also have resources for confronting the disease that could not have been imagined decades ago, including: genetic sequences, sophisticated blood tests, and powerful computer models.

Right now, much of the world is experiencing an eerie calm before a likely storm. Regions that have taken strong actions to reduce transmission – like closing schools, restricting movement, and limiting commerce – have at least bought some time, hopefully time that will be used to make good decisions.

It’s not clear how much time we have bought, though. Strong actions take a heavy toll on the economy, on the fabric of society, and on people’s mental well-being. We must use the calm before the storm to plan how to deal with various scenarios that may unfold in the future. To do that well, we first need to know more about COVID-19. We need to learn as much as we can now, before it is necessary to make really hard decisions about balancing disease control, the economy and our social fabric.

So what do we need to know? Things we could study to enable better decisions after the calm include patterns of immunity, and time distributions of illness and infectiousness.

Perhaps the most important gap in our knowledge has to do with patterns of acquired immunity. No matter what happens in the short or medium-term, COVID-19 will not cease to pose a grave threat until most of us have acquired immunity, either via vaccination or via natural infection. Everyone who has had a SARS-CoV-2 infection, however mild or severe, will show signs of acquired immunity in a simple blood test. Because we still don’t know much about asymptomatic infections with SARS-CoV-2 – neither how common they are nor how dangerous – we don’t really have a good idea of how much immunity has already been acquired, nor what proportion of all the people who will eventually be infected are likely to have serious outcomes.

Stratified “sero-surveys”, aimed at getting a population-level view of who has been infected in places around the globe – particularly places with large outbreaks, like Hubei or Lombardy – can provide a tremendous amount of detail about the age-specific attack rate (probability of getting infected) as well as the consequences of infection. Much attention has focused on the number of deaths, hospitalizations, and severe outcomes per thousand “cases”. But infections with SARS-CoV-2 don’t always count as cases, and case definitions are always tricky to apply, and typically wind up meaning different things at different times and places. What we really want to know is the number of negative outcomes per infection, and the best operational definition of having been infected in this context is having acquired immunity. People with acquired immunity to COVID-19 have fought off a SARS-CoV-2 infection successfully, which means that they’ve been exposed to SARS-CoV-2, may have transmitted it, and are very unlikely to be in danger of serious consequences from future SARS-CoV-2 infection.

Another key gap is information about time profiles of human illness and transmission after infection. What proportion of people are able to transmit before they show clear symptoms? What proportion transmit without ever showing symptoms? How long do diseases generations take on average, and how variable are they? Sero-surveys can help with these questions as well. More can be learned about time profiles by detailed contact tracing: who do we think infected whom, who showed symptoms when? And more again by studying patients symptoms and viral profiles in detail, including by colllecting nasal swabs from people in isolation.

It is likely that we have difficult decisions ahead. We can act now to make sure that we make them with better information.

Hat tip

Thanks to Daniel Sang Woo Park, Joshua Weitz and Bill Hanage for suggestions and clarifications.