Times Square, nearly empty on the first day of the New York Stay at Home order. | Photo by Lev Radin/Pacific Press/LightRocket via Getty Images
Why the coronavirus lag time makes it hard to figure out whether social distancing measures are working.
Now that most states have shut down temporarily to combat the coronavirus, our greatest enemy is our own impatience.
The measures taken over the past two weeks in 37 states and parts of eight more hurt. They kill off businesses that are critical to the livelihoods of millions of people. They throw millions of Americans out of work and off their health insurance.
And for a little while longer, it won’t even be clear if they’re doing anything.
“We have sacrificed so much already, but it feels like nothing is working. That’s because it takes weeks to see results. We must stay committed and trust that the social distancing we are enduring now will save thousands of lives,” Dr. Caitlin Rivers at the Johns Hopkins Center for Health Security wrote on March 23.
Here’s why it will take so long for our sacrifices to have visible effects. The thing we’re trying to reduce — new coronavirus infections — is invisible at first. It takes between two and 14 days for a newly infected person to start showing symptoms. After symptoms begin to show, it can take more than a week for them to be eligible for testing (many people are not eligible at all). And then, thanks to backlogs in testing availability, it can take days for them to learn they tested positive.
As a result, every positive test today reflects infections that occurred, on average, a few weeks ago. If shelter-in-place orders worked, we should just start to see their effects this week in the states that acted soonest, and it will take weeks more to see their effects nationwide.
New York’s numbers appear to be falling, but it’s too soon to say that the city has rounded a corner thanks to its March 20 stay-home order — these infections being detected by tests now would nearly all have occurred before the order happened.
In other words, once you shut down your city to fight the virus, you might even see things get worse before they get better. There’s a very real risk that that will discourage us and prompt us to give up. We shouldn’t.
Lag time, explained
Let’s say we lock down a city to prevent the further spread of the coronavirus, and the lockdown works very well, as the one in Wuhan seems to have. Where previously the average infected person spread the disease to two or three other people, after the lockdown in Wuhan it is estimated that each infected person only spread the disease to .32 people.
If we accomplished the same thing, the number of new infections would start declining right away. But here’s the thing: The number of new positive tests? That would keep growing.
Here’s why: Symptoms of the coronavirus typically take some time to start showing after infection. The average incubation period is five to six days but “may range from 2-14 days,” according to the Centers for Disease Control and Prevention. So keep that in mind when thinking about people who are starting to develop symptoms now: Those were people who were exposed to the coronavirus about a week ago, in some cases even longer.
Now, most people cannot get a test as soon as they start to show symptoms. In some parts of the country, in fact, it’s impossible to get a test until you are hospitalized, unless you are a health care worker or are otherwise considered high risk.
The coronavirus has a long course of illness — people are usually sick for a while before they get sick enough to need hospitalization. In one study from China, the average time from symptom onset to hospitalization was nine days, and in another study, difficulty breathing usually surfaced five to 13 days after symptoms started.
That means that, in the parts of the US where most patients aren’t tested for the coronavirus until they are hospitalized, a patient may be tested 14 days after they were infected — five days for the average patient to show symptoms and nine days after that to end up in the hospital.
Finally, tests don’t always get results right away. There are some testing sites that advertise same-day turnaround, but other labs are backlogged, and in many parts of the country results take longer. Rand Paul’s positive test for coronavirus reportedly took six days to come back.
Add it all up, and it will take at least two weeks, likely longer, for a coronavirus infection to be reported in official statistics.
That means that it will look — for weeks — like the measures we’re taking now like self-quarantines, school closures, and social distancing are not working, even if they are. And if they’re not actually working? Same thing — we won’t know just yet. The data out of the San Francisco Bay Area (which issued its shelter-in-place order on March 17) and New York (March 20) over the next week will be critical to understanding how stay-home orders work. We should watch those numbers closely to understand our next steps.
Experience from other countries
On January 23, China locked down the city of Wuhan amid a growing epidemic. On that date, the country had 830 cases. Much of the rest of Hubei province locked down later that week, and then restrictions were imposed on the whole country.
For the next few weeks, things got worse. By February 13, the country had 63,851 cases — most of them in Wuhan’s Hubei province. But things were starting to turn around. That day, China reported 5,090 new cases — more than on any previous day (except the day of a one-time change in the types of cases reported). On February 14, however, officials reported 2,641 new cases. On the 15th, 2,008. New case numbers kept declining from there; China now reports no new local transmission cases (though they could be either missing some or hiding some, experts don’t think they’re hiding a full-blown epidemic.)
Take a look at that gap again: The peak of new cases came fully three weeks after the lockdown started. In Italy, it’s starting to look like the same dynamic is at work.
“These big social distancing measures take time to work,” Johns Hopkins Center for Health Security director Tom Inglesby tweeted March 23. “The impact of big interventions in Wuhan China took about 3 wks to start to reverse things. And then everyday after the situation got better.”
Should the US expect the same thing?
Not necessarily. Some things are definitely different. The US implemented social distancing measures more gradually than Wuhan, which escalated quickly from few restrictions to a full lockdown. No place in the US is employing measures as strict as Wuhan’s. China separated symptomatic people from their families and locked people in their houses. Enforcement of stay-at-home orders has, so far, been lax in the US.
But if our measures suffice to reduce transmission so that each new patient infects less than one person, it wouldn’t be surprising if the overall trajectory of the disease in a city like New York looks somewhat like China’s, with cases peaking three weeks after the measures went into place. That would put us into early April. (There are some signs that cases are flattening already, only a week after the stay-home order, but it remains to be seen if these gains will last, and it’s probably a mistake this early to attribute them primarily to the stay-home order.)
Graphs like this one are cited as evidence that the Bay Area’s shelter-in-place policy has worked, when they largely reflect infections from before shelter-in-place went into effect.
In other parts of the country, where those measures haven’t yet been put into place, the peak is even farther off.
“We know when this started. We can get a good estimate on when it will end,” former FDA Commissioner Scott Gottlieb argued on March 24, projecting a peak in New York in two to three weeks.
It’s difficult to be patient and wait for that. The costs of social distancing measures — upended routines, shuttered stores and restaurants, lost jobs, vulnerable people put at risk — are apparent already, while the benefits aren’t. Social distancing measures save lives, but the lives they save are a month away while the lives they throw into turmoil are visible right now.
Unsurprisingly, there is overwhelming pressure to identify results from these measures right now. And preliminary data that looks good is not hard to find. Articles have proliferated comparing states that implemented social distancing to states that didn’t, or declaring that early data shows California’s stay-at-home order is working.
It’s tempting to draw those conclusions. But it’s really too early to say any of those things with certainty. The numbers published today reflect measures taken in the first weeks of March — before the stay-home orders began. They don’t reflect the effects of the lockdowns, or even, in many states, the school closures. The states that are doing well cannot credit their success to measures they’ve taken in the last few weeks, and the states that have imposed such measures shouldn’t expect results yet.
What’s important to remember is that if the numbers suddenly get worse in a state that recently implemented stay-at-home measures, it won’t prove that stay-at-home doesn’t work. The confirmed cases we are seeing today are largely people infected a few weeks ago, and the data on how well our current measures are working won’t be available for a while.
“For a while it will feel like nothing is working but it takes time,” Rivers wrote March 23.
When things are changing as rapidly as they are with this crisis — and when families are struggling with lost jobs, wiped retirement accounts, sick loved ones, and closed schools — the few weeks we’ll need to wait before case numbers hopefully start declining feel like an eternity. But the best thing we can do is help our cities develop a plan for the next stage of virus response, support essential workers, stay home, and wait.
Sign up for the Future Perfect newsletter and we’ll send you a roundup of ideas and solutions for tackling the world’s biggest challenges — and how to get better at doing good.
Future Perfect is funded in part by individual contributions, grants, and sponsorships. Learn more here.