Dozens of states across the US are pinning their hopes on contact tracing to control the spread of the coronavirus and enable regions to reopen without sparking major resurgences of the outbreak.

Alaska, California, Massachusetts, New York, and others are collectively hiring and training tens of thousands of people to interview infected patients, identify people they may have exposed, and convince everyone at risk to stay away from others for several weeks.

Contact tracing is a proven tool in containing outbreaks of highly infectious diseases. But this particular virus could pose significant challenges to tracing programs in the US, based on new studies and emerging evidence from initial efforts. Stubbornly high new infection levels in some areas, the continued shortage of tests, and American attitudes toward privacy could all hamstring the effectiveness of such programs.

Driving down infection rates

The chief challenge with this coronavirus is its potential to spread exponentially: absent containment measures, every infected person on average will infect two or three others, according to most estimates (although some studies find it could be higher).

The goal of contact tracing, as well as social distancing, is to push down the number of people each infected person infects, creating an “effective reproduction number,” or R_e, of 1 or less. At that point the number of new cases is flat or falling.

But contact tracers have to reach a significant portion of cases and contacts to really move those numbers.

A team in any given region would have to detect at least half of new symptomatic cases, and reach at least half the people they were in close contact with and encourage them to stay away from others, in order to reduce the transmission rate by 10% or more, according to a new model. (The work was published as a preprint on MedRxiv on May 8 but hasn’t been peer-reviewed yet.)

If they successfully detected 90% of symptomatic cases and reached 90% of their contacts—and tested all of them regardless of whether they had symptoms—it could reduce transmissions by more than 45%, the researchers found.

In other words, if social distancing in a given region had reduced infections per person from 2.6 to 1, this level of contact tracing could push it down to .55. Or the region could ease distancing measures by about half and keep infection levels constant.

“It gives us some room to be targeted and strategic in terms of the sorts of restrictions we have on business and commerce and social interactions,” says Joshua Salomon, a professor of medicine at Stanford and coauthor of the study.

Can we hit those kinds of figures? Salomon thinks it’s possible, but he adds that most of the nation doesn’t have the well-trained workers and data systems in place to achieve anything on that level yet.

Amassing armies

The success of contact tracing will depend on how big the teams are, how many new cases develop, and how readily people respond in any given community.

Reaching 90% of contacts, for instance, will be especially difficult in states and regions still grappling with lots of new infections. Take Massachusetts, which put to work a 1,000-person contact tracing task force at the beginning of the month. But new confirmed cases in the state are still generally exceeding 1,000 daily, and nearly reached 1,700 on Thursday, so every tracer on the team will need to track down and convince some multiple of that number to stay away from others every day. While shelter-in-place rules are in effect, that multiple may be only two or three people. But as regions relax social distancing measures, the average number of contacts for infected patients could rise to closer to 20.

NPR reported that 44 states and the District of Columbia now plan to bolster their contact tracing teams, collectively increasing them from about 11,000 today to more than 66,000 in the weeks ahead.

But that’s likely not going to be enough. The National Association of County & City Health Officials estimates that US tracing efforts will require 30 professionals for every 100,000 people (or more than 98,000 people nationwide).

Only seven states have plans that would reach that target, including California, New York, and Illinois, NPR found. Only one, North Dakota, meets it currently.

A bipartisan group of prominent health experts and public officials—including Bob Kocher, a former special assistant to President Barack Obama on health policy, and former Republican Senate Majority Leader Bill Frist—argue that the nation will need to go further than that. They called on Congress to set up a 180,000-person contact tracing workforce that would cost the federal government some $12 billion.

Technology can certainly supplement human contact tracing. Smartphone apps that flag when someone may have been in close contact with an infected person helped China, which required citizens in many cities to download the software, to flatten the curve of its outbreak. Similarly, South Korean officials have made use of credit card information, surveillance footage and GPS data to track the recent whereabouts of patients who test positive, as well as smart phone apps to ensure they remain in quarantine.

But proximity-tracking apps, at least, need to be very widely used in order to make a meaningful difference. So there are serious doubts about how effective they could be in the US, given heightened cultural concerns about privacy and the fact that the government isn’t forcing people to use them.

Move fast and test things

A study published on May 1 in JAMA Internal Medicine, tracking the first 100 cases in Taiwan, found that people are most infectious before and within five days of the onset of symptoms. That adds to a growing body of evidence that people with minimal or no warning signs like fevers and coughs are a major vector of the disease.

That underscores the critical importance of contact tracing. The very goal is to identify people who don’t know they’re infected and encourage them to quarantine themselves before they unwittingly infect others. But it’s hard to identify and trace all the cases out there if people aren’t sick enough to know they should get tested, and it means contact tracers need to move incredibly fast to get to people before they’re already spreading the virus.

“I say you need to find people and isolate them within four days of exposure, if you’re going to make a dent,” says George Rutherford, a professor of epidemiology at the University of California, San Francisco, and principal investigator on California’s contact tracing program. “It’s probably even three.”

The potential for people to spread the disease before displaying symptoms also underscores the importance of making tests much more widely available. Given the shortages of supplies, protective equipment, trained personnel, and processing capacity, many regions are still offering tests only for people who have symptoms or are frontline health workers.

But if areas can build the capacity to test all the close contacts of infected people, even if they haven’t developed symptoms, it could increase the effectiveness of contact tracing programs by as much as 2.2 times, Salomon and his coauthors found.

That’s because if an asymptomatic person does test positive, it triggers additional efforts to contact and quarantine everyone that person may have exposed. Also, researchers suspect, people are more likely to adhere to requests to stay away from others if they know they’re infected than if they’re simply told they may have been exposed.

Last month, researchers at the Harvard Global Health Institute estimated that the US would need to conduct at least half a million tests per day to capture asymptomatic cases and safely reopen the economy. Now they think that figure is closer to 900,000. Meanwhile, the daily average over the last week has been about a third of that level, according to the COVID Tracking Project.

The American psyche

Successful contact tracing efforts also require people to accept calls and heed advice from complete strangers.

Unfortunately, years of robocalls and telemarketing have conditioned many Americans to ignore calls from numbers they don’t recognize. Jana De Brauwere, a program manager with the San Francisco Public Library who is working with the city’s contact tracing task force, says that at least half the people she calls simply don’t answer. Others hang up once she starts asking for personal information, like addresses and dates of birth.

UCSF’s Rutherford has said there’s an added challenge for regions with large populations of immigrants or undocumented residents, where people may be fearful of interacting or sharing information with public officials. San Francisco’s contact tracers are finding that about 40% of potentially exposed contacts are Spanish-only speakers, many of them in crowded living situations.

Even if contacts do take the call and stay on the line, there’s the separate question of whether they’ll follow the advice to get tested or voluntarily place themselves in quarantine.

De Brauwere says that all she can do is recommend these steps, and offer support to help people take them. She can connect them, for instance, with city workers who will deliver food or medicine, or even locate shelter in extreme situations.

Kocher, a fellow at the University of Southern California’s Schaeffer Center for Health Policy and Economics, adds that we’re often asking people to miss work for several weeks. Some people can’t afford that, and some will fear losing their jobs. So if we expect people to comply, we may have to provide additional incentives, including money to pay their bills, he says.

There are other reasons to suspect these requests won’t go over well among certain people and in certain parts of the nation, particularly as public opinion about the dangerousness of the disease and the appropriateness of government interventions becomes increasingly politicized.

In a Twitter thread earlier this week, Keith Humphreys, a professor of psychiatry and behavioral sciences at Stanford, argued that public health officials are underestimating how much US attitudes toward government authority could undermine national testing and tracing programs.

Americans have already defied the orders of health officials in several prominent incidents, including assaults on store workers who asked people to wear masks, armed demonstrators protesting stay-at-home restrictions, and businesses that have reopened before their local government gave the go-ahead.

Public health orders, Humphreys notes, work only when there’s a public that will abide by them.

Update: This piece was updated to clarify how South Korea has used digital tools and other methods to control the outbreak in South Korea.