Ebola-hacking lessons for coronavirus fighters (Q&A)

Ebola was running rampant across Sierra Leone in 2014 when Emerson Tan, a hacker with the IT and nongovernmental development organization NetHope, was sent there with a vague mandate to “improve things.” He and a team of local hackers did better than just ameliorate the situation; their work was instrumental in eliminating Ebola in the country.

Six years later, Tan says, he’s frustrated that the lessons his team learned in Sierra Leone are largely being ignored, as the coronavirus that causes Covid-19 roils the globe.

“This isn’t just a medical problem. It’s a medical problem in so much as it’s an infectious disease borne by a pathogen,” Tan said during a call from his home near London. “But actually it’s a test of the entire social fabric. If your country has a strong social safety net, it will have fewer deaths than a country without a good social safety net. Sierra Leone had no social safety net, and we’re finding out now which countries have good social safety nets by their death counts.”



READ MORE CORONAVIRUS COVERAGE

Many of the problems that Sierra Leone and other West African countries faced during the Ebola crisis, which stretched into 2016, will sound familiar to anybody who’s keeping up with Covid-19 headlines: A lack of personal protective equipment (PPE) for medical practitioners on the front lines of the epidemic; a shortage of life-saving medical equipment such as ventilators for patients; and confusion, unwillingness, and uncertainty among the general population to take the steps necessary to protect themselves from infection.

When Tan was in Sierra Leone, one of the teams he was working with focused on creating and implementing a payment system built on open-source software designed to ensure that medical professionals on the front lines of the epidemic could get paid.

The virus will exploit every disconnect. The lesson of Ebola is unity of action. Where you have people who don’t believe that it’s real is where the virus goes on a rampage.

That may seem tangential at best to solving a medical crisis, but regular, timely payments helped build the confidence of doctors and nurses that the highly dangerous nature of their work was being taken seriously. It eliminated their need to ask patients for financial help, and kept them from staging work strikes. As confidence among medical practitioners grew, so did that of patients, many of whom were reluctant to seek aid in clinics or hospitals, and often relied on in-home, traditional medicines unable to combat Ebola.

Ultimately, the World Health Organization found that Sierra Leone’s health care workers were between 21 and 32 times more likely to become infected with Ebola than non-health care workers. Ebola would claim the lives of 21 percent of all health care workers and 41 percent of those infected across the country during the epidemic.

The end game for the coronavirus is nowhere in sight. And while Tan castigates authorities for not responding quickly enough to the pandemic, he says there’s still time to look to past epidemics to minimize the Covid-19 mortality rate and keep people from making deadly mistakes.

At the time of publication, it’s hard to say exactly what Covid-19’s mortality rate has been. Worldwide, the Johns Hopkins University’s coronavirus tracker estimates that 5.8 percent of people infected with Covid-19 have died. But because few governments are disclosing numbers of deaths of citizens who may have been infected with coronavirus but were not tested for it, that figure could be way off.

Illegal economic migrants excluded from the social safety net, excluded from the health care system, are the perfect reservoir for the virus. They don’t appear in our surveillance. The minute you leave people out because they’re too poor, or black, or Asian, or because of their immigration status, you give the virus a place to hang out, and then it’ll come back. The minute you cut people out of these systems, you prolong the epidemic.

In short, Tan says what should be a familiar refrain by now: Wash your hands early and often (Dr. Anthony Fauci of the National Institutes of Health says he washes his hands 100 times a day), wear face masks whenever you interact with people outside your home (if only because wearing a mask reminds you not to touch your face), and stay at home as much as possible to avoid interacting with people who may be infected—regardless of whether they’re displaying symptoms. What follows is an edited transcript of our conversation.

Q: What’s the No. 1 lesson you learned in Sierra Leone that you think is being ignored during the coronavirus pandemic?

We had to build a makeshift social safety net. Because of shut-down supply lines and an  aggressive quarantine, we had to start feeding people. We had to arrange for food shipments. In the United States and other countries now, the logistics chain is largely intact. But because of the restriction orders, it’s starting to fray at the seams.

Take California lettuce, grown in the Central Valley. Who cleans it? Who packs it? Who ships it? The most important people in this are on minimum wage, unable to cope, and because so much of the U.S. food economy is done by undocumented people, they have a strong interest in not participating in social safety measures.

Emerson Tan stands in front of the last quarantine center in in Sierra Leone to shut down. Photo courtesy Emerson Tan.

Illegal economic migrants excluded from the social safety net, excluded from the health care system, are the perfect reservoir for the virus. They don’t appear in our surveillance. The minute you leave people out because they’re too poor, or black, or Asian, or because of their immigration status, you give the virus a place to hang out, and then it’ll come back. The minute you cut people out of these systems, you prolong the epidemic.

You want to keep this short, you create a universal system for everybody. Once there’s a vaccine, there should be only one criteria: Do you have a pulse? Everybody gets it. For everyone who does not get it, that provides somewhere for the virus to hang out, and then we’re back to square one.

We need to protect these economic migrants to protect the majority of the population. They are also the ones you can absolutely not afford to not work because then people in cities starve. The communities in Sierra Leone that did best were subsistence farmers. You’re more or less self-sufficient, and the second- and third-order effects affect you a lot less. But once you put a city into lockdown, you also have to be careful about how you restart everything.

Before we get to restarting social interactions and the economy, can you talk about how long it took Sierra Leone to stop its Ebola epidemic? What can the experience tell us about the coronavirus pandemic?

Ebola took almost a year to stamp out completely, after the main epidemic was over. You’d see a case out of nowhere, which meant there was an unseen chain of transmission. You’d see a data point, contact-trace it back to a spreading event, then isolate the people at and around that spreading event.

Covid-19 is really hard. We might see a spreading event, then a neighborhood or citywide period of lockdown. But we can’t always go back to a spreading event. In Sierra Leone, it would be funerals. We’d trace an Ebola infection back to one person that had died, and would realize that everybody at the funeral had been infected.

Covid-19 is so spready that we might not be able to do that. Instead of quarantining everybody at a church, you might have to quarantine a whole neighborhood. For a megachurch [such as one in France that spread the coronavirus far beyond its community in February and led to scores of deaths], forget it. We might have to ban traditional funeral practices.

For many rural Sierra Leoneans, the world is a much more rich and magical place. It’s full of magic and spirits and traditional beliefs, a psychic cosmography that’s completely different from the Western scientific viewpoint. They believe that if you die without ritual washing, then your power and magical essence leaves your family. The ritual washing helped the virus spread.

We’ll surveil everybody, and just to play it safe, we’ll keep the data forever. At this stage of the pandemic, that’s a terrible idea.

If your world view doesn’t accord well with reality, then the virus will exploit that. And the same thing is happening here. Holding a belief that the virus somehow distinguishes between believers and nonbelievers, or is somehow a Democrat hoax, influences your behavior and makes you far more unsafe.

We might have to park high-risk activities for a long while. We might see small church services allowed but megachurches banned. Little League might be OK, but Major League Baseball, with tens of thousands of attendees, wouldn’t be.

In keeping communities as safe as possible from Covid-19, isn’t knowing who has it and who spread it extremely important?

People get very hung up on ideas of being able to go back in history to see where you’ve been, or being able to leverage the surveillance technologies they already have. We’ll surveil everybody, and just to play it safe, we’ll keep the data forever. At this stage of the pandemic, that’s a terrible idea.

You should be able to say some sensible things about what your data policies should be at this point. If you can’t figure out at this point what’s worth keeping and what you can throw away, you’re probably doing it wrong—or you haven’t thought it through particularly well. There’s no reason that you can’t implement this in a way that’s fit for purpose and protects people’s privacy. Those aren’t completely exclusive things.

I would be very cautious about people that say it’s impossible to design systems that don’t preserve people’s liberty, or that there’s a feeling that you have to submit people to some kind of panopticon to keep them safe. Once you establish the parameters of the disease, which is pretty clearly defined at this point, there’s no reason to infringe on people’s rights.

You’re talking about deleting data once it’s no longer needed. How can consumers know that location data that’s been collected for legitimate coronavirus prevention reasons has been dumped?

It’s very difficult to validate that data has been thrown away. I would struggle to give positive validation that your data had been binned after a certain exercise. But there’s no reason that you can’t give people legal assurances that their data has been ditched. There’s no reason that you can’t build things in such a way that they need to be like that.

There’s a great temptation to try and collect it all, and then try to figure out what to do with it all.

Pokemon intelligence is what the Pentagon used to do—they gotta collect it all! But there’s so much unusable data that it becomes irrelevant. It loses all of its predictive value. The classic example of this is location data.

What does that mean from an engineering perspective? Simply not collecting information that’s not relevant, and ensuring that data is chucked when it’s not relevant anymore.

Let’s say I collect anonymized location data. I’m interested in where you are and where you’re going. To stop the spread of the coronavirus, what I’m interested in are events of people in high-density areas for more than a half-hour window. Churches, gigs, baseball games, concerts. But I’m not interested in whether you went to synagogue this week. What is interesting is that the synagogue is open when it should’ve been closed, and that groups of people went to it.

Now I can take an enforcement action. I’m not interested in arresting people; I’m interested in shutting it down. Penalizing individuals is not useful. Enforcement actions aimed at individuals are risky. Enforcement action aimed at institutions is worthwhile.

Let’s talk about contact tracing and the rise of aggressive geolocating of people as part of the effort to stamp out the spread of the coronavirus. I know you’re working on a way to make this process more protective of everyone’s privacy. What does that look like?

I’m working with the Covid Alliance to create some markers as to what’s technologically useful.

One marker is that when you’re immune to something, we don’t care about you. We can toss your data. Immune means that you’ve had an antibody test and that test has come back positive.

That doesn’t mean a positive antigen test, or another test. The antibody test means that you can not be infected again for the foreseeable future. We haven’t seen any conclusive data that people can be reinfected.

[An antibody test can identify patients who have had the coronavirus but were asymptomatic or developed symptoms that weren’t diagnosed. An antigen test can only detect if the virus itself is currently present in a patient.]

Another marker is that you are infectious from a period of 7 to 10 days after exposure through to about a week to two weeks after the end of your symptoms. What that means is that that’s the only window of time where I care about what you’ve done. What this means is that I can temporarily bound the window of surveillance that I actually need to surveil you for.

Unless you have real-time data, you’re always two weeks behind—which is a terrible place to be. It means the disease is doing the leading. We don’t have the initiative.

Here’s a hypothetical example: You report symptoms, go to hospital, and the test comes back positive. At this point, I care about where you’ve been for the past 10 days. I care about you obeying the quarantine—no free hugs for everybody. I care about the notion that at the end of this, the disease has seroconverted in you.

What does that mean from an engineering perspective? Simply not collecting information that’s not relevant, and ensuring that data is chucked when it’s not relevant anymore. If you’re an ex-smoker, 69 years old, you’re high-risk. If I have a rolling window of surveillance, it works perfectly well for you because your data gets tossed.

This means we have to be much more selective about how we do enforcement. It allows you to be more fine-grained about enforcement. Arresting a homeless black guy doesn’t do much to prevent transmission. Shutting down the illegally open megachurch does.

Enforcement against three people having a picnic doesn’t do much. Chaining the gate shut to the park does. Discouraging large groups is good instead of shutdowns, but only if people are following the orders.

Unless you have real-time data, you’re always two weeks behind—which is a terrible place to be. It means the disease is doing the leading. We don’t have the initiative.

As we’ve seen in Sierra Leone and in places impacted by other epidemics, we will eventually be able to restart the economy. What does that look like?

The American Enterprise Institute plan calls for a soft start. You only restart general movement and the rest of it when you have the infrastructure in place to do contact tracing and aggressive isolation.

You only turn things back on when you have enough data-based surveillance to know who’s been in contact with whom. You won’t catch everything. But when you have a case that actually spits out a data point, you can trace the transmission and kill the hot spot.

The aim of the game here is to accelerate the downslope, and also ensure that as vulnerable populations open up, it doesn’t take off again. You don’t give the virus any room to come back. You isolate everyone around it, and work backwards. That’s the only way you can restart things.

Otherwise, you get a second wave of infections, and you’ve opened up a new vulnerable population. You only turn things back on when you have enough data-based surveillance to know who’s been in contact with whom. You won’t catch everything. But when you have a case that actually spits out a data point, you can trace the transmission and kill the hot spot.

Despite being the third-most populous country, the United States has more than double the number of reported coronavirus cases as Spain, the country with the second-most cases. What other lessons from Ebola-infected Sierra Leone can the U.S. still use in its struggle to contain the coronavirus?

Nobody’s really interested in Sierra Leone because it’s poor, but there are some clear analogues between the U.S. and Sierra Leone. There’s the chronic distrust of central government, and that massively impedes your ability to get things done.

Trust is critical because you’re asking people to take steps to fight a threat that they can’t see. If it works, they will never experience it. For a lot of people, the coronavirus won’t feel real until somebody they know gets it.

What swung it for Donald Trump were the hospitals in New York that look like those from Italy. Corridors of people, intubated, dying, in body bags.

All the normal sources of social cohesion in the U.S. are fragmented. The virus will exploit every disconnect. The lesson of Ebola is unity of action. Where you have people who don’t believe that it’s real is where the virus goes on a rampage.

Think of the virus as an alien invader. It knows nothing about people, about our social structures, if you’re president, prince, or pauper. You’re a walking bag of meat that’s a good portable home, a resource to be consumed. It doesn’t even care if you live or die.

In Sierra Leone, cases came from out of nowhere, and we’d inevitably discover that there was a surveillance failure because people have a reason to keep something secret. It’s far better to accord everybody a minimum safety net or minimum set of rights so that everybody is incentivized to play along.

Individualism doesn’t win wars. You can recover from being broke. The only thing you can’t bounce back from is being dead.