An unmitigated disaster. That’s how Lawrence Gostin of Georgetown University describes the handling of the Diamond Princess.
For two weeks, the British cruise liner sat quarantined off Japan’s coast, and roughly 3,700 passengers and crew were housed inside a giant incubator for the novel coronavirus, which causes the disease officially known as COVID-19. What started as a manageable group of 10 infections on February 4 ballooned into more than 600 cases by Wednesday, when everyone on board was finally allowed to disembark.
“What should have happened is, they [the passengers and crew] should have been disembarked from the boat and placed in isolation or medical quarantines,” says Gostin, a professor who is also director of the World Health Organization Collaborating Center on National and Global Health Law.
For him and other health academics, the Diamond Princess saga speaks to the imperfect legacy of travel restrictions and disease screening. For centuries, public officials have combed travelers for signs of disease, whether it be for cholera in the 1800s or Ebola crises in the last decade, and cordoned them off.
Yet time and again, history has shown that stemming the spread of an infectious disease becomes more difficult when these control methods are deployed inappropriately or unevenly. Border closures protected some farming villages during Spanish Influenza in 1918 and 1919, but they also kept countries like Portugal from obtaining health resources. Meanwhile, the primary spreaders of the disease—troops in World War I—easily crossed borders.
Back then, people barely understood how diseases moved among human populations, but today’s epidemiologists have found that the same overarching rule applies to modern outbreaks: Borders are leaky no matter what you do. Now, in the face of COVID-19, these scientists are creating tools that can ostensibly predict the weak spots—and judge whether tactics like the lockdown of China’s Hubei Province or the Diamond Princess blockade are worth the consequences.
Did the Wuhan lockdown work?
Viruses are inherently clandestine, and the one that causes COVID-19 is no exception. As of Friday, this outbreak has infected 76,775 and killed 2,248. Early evidence shows that the novel coronavirus takes about five days to manifest symptoms in mild and severe cases, but patients are contagious before these symptoms arise.
COVID-19 readily establishes itself not only in the lungs, but also in the upper respiratory tract, meaning the nose and throat. A study published Wednesday in the journal Science reports that the novel coronavirus’ affinity for entering human cells is 10 to 20 times higher than that of other coronaviruses. These habits mean the new coronavirus can more easily hitch a ride via coughs and sneezes, and they may explain why people become contagious before displaying full-blown symptoms.
Those details matter, because the timing of when people become contagious can help gauge if changes in travel policy—such as China’s transportation bans and international airport screening—are truly keeping the novel coronavirus from crossing borders.
Matteo Chinazzi, a network scientist at Northeastern University, has co-developed a way to judge the effectiveness of COVID-19 travel bans, both within and against mainland China. The project hinges on the Wuhan lockdown on January 23, when China restricted movements within the city of 11 million. The model charts how global and domestic commuters flowed before and after this turning point in the outbreak, combining high-resolution population data and disease-tracking algorithms tailored to the ongoing epidemic.
“Our model is constantly updated and recalibrated using new information and the updated interventions that different countries are implementing,” Chinazzi says. The result offers a readout on the effectiveness of these travel restrictions, and so far, the verdict is mixed.
First, China’s lockdown of Wuhan likely arrived too late; the model predicts that the novel coronavirus had already established footholds in other major Chinese cities by January 23. That means mainland China was likely already exporting coronavirus cases through other travel hubs, with the model pointing to Shanghai, Beijing, Shenzhen, Guangzhou, and Kunming as the highest-ranked sources.
The lab’s predictions may also explain why the outbreak continues to thrive in certain places, like Japan and South Korea. Rather than lower the odds of transmission, the Wuhan lockdown increased the risk for coronavirus importations for these two countries, which only issued partial transportation bans. Travelers from Wuhan and Hubei Province were blocked, while visitors from other parts of mainland China could still enter.
Overall, the model suggests that the Wuhan lockdown only delayed the global progression of the epidemic by three to five days. Such delays aren’t worthless, given that they provide precious time to coordinate a response. But their limited effectiveness should be no surprise, considering research conducted on the last two decades’ worth of outbreaks.
“There is not a lot of evidence that a travel ban would completely eliminate the risk of an infectious disease spreading in the long term,” says Nicole Errett, a lecturer at University of Washington School of Public Health and a former special assistant at the U.S. Department of Health and Human Services.
In the latest issue of the Journal of Emergency Management, Errett and two colleagues reviewed past travel bans implemented for Ebola and SARS, and they reported that most were only effective in the short term. Similar investigations for influenza found that travel bans could delay the spread of epidemics by one week to two months, but the overall incidence of the disease only dropped by 3 percent.
Does airport screening help?
Meanwhile, the methods for spotting the disease among people on the move aren’t foolproof. Take, for example, the temperature guns that you see pointed at people’s foreheads at airport customs and border checkpoints. On average, those devices are only 70 percent effective at detecting fevers, meaning about one of every four people with elevated body temperatures goes unnoticed.
“Travelers screening is not some kind of firewall that will absolutely protect from having cases imported into whatever area you're trying to defend,” says Jamie Lloyd-Smith, an infectious disease ecologist at UCLA. “This is not because [the screening] is being done poorly, and it’s not because the people who are in charge are being lazy.”
A man walks by windows at Hong Kong International Airport on February 14. PHOTOGRAPH BY LAM YIK FEI, THE NEW YORK TIMES
Lloyd-Smith and other mathematicians are assessing the strong and weak points in travel screening as it pertains to the COVID-19 outbreak. Their latest work is inspired by a study they published in 2015, which built a model that systematically estimated the performance of traveler screening programs during outbreaks of SARS, MERS, influenza, and Ebola.
As in this earlier work, the new model for COVID-19 includes basic factors like the failure rate for thermometer guns or how easily the virus moves between people. But it also accounts for more subtle variables, like how many people might transmit the virus before their symptoms arise, or how often people accurately report their symptoms on screening questionnaires distributed at airport customs.
“It's pretty evident from past outbreaks that people aren't always honest about risky exposures,” Lloyd-Smith says. “Based on the past data we had, it looked like one in four passengers accurately and honestly reported the risky exposures they had.”
Compute all these variables, and their model estimates that enhanced screening is at best catching 50 percent of infected air travelers and at worst just 20 percent, primarily because COVID-19 symptoms are so latent.
“Based on what we understand about this virus, something on the order of half of people are fundamentally not detectable [during screening],” Lloyd-Smith says. While the results are currently being peer-reviewed at the journal eLife, they mirror what preliminary results from a separate research group at the London School of Hygiene and Tropical Medicine predict for air traveler screening with COVID-19.
“We'd probably only catch about 45 percent of infected travelers using exit screening,” says Samuel Clifford, an epidemiologist at the London School’s Center for Mathematical Modeling of Infectious Diseases. “Out of the remaining 55 percent of people who aren't caught, we can catch a few more on entry. You've got 42 percent of the people [with COVID-19] still making it into the country.”
Singapore, the bellwether
While such models can offer an indication of what might be happening, you can see these dynamics play out in real-time by setting your sights on one nation: Singapore.
Singapore boasts one of the highest-regarded health care systems on the planet, thanks to public funding, low costs for treatment, and an abundance of doctors, nurses, and other health professionals. This workforce pays off during an outbreak, because it means that Singapore can catch cases the moment patients arrive on their shores.
“Singapore had such a remarkable record during SARS with following cases, and it seems to have a high ratio of detection this time around,” says Marc Lipsitch, an epidemiologist at Harvard T.H. Chan School of Public Health. He thinks tracking the COVID-19 situation in Singapore can help gauge how the outbreak might evolve in other countries, especially in areas with developed health care systems or places that receive high volumes of travelers from China.
The United States fits both aspects of this billing, given that it receives as many travelers from China as Singapore, roughly three million a year.
Based on preliminary modeling in Lipsitch’s lab, Singapore ranked first out of 191 countries in disease surveillance during the first weeks of COVID-19. His team suspects that Singapore's methods are exceptionally sensitive to new cases, and can thus serve as a standard of comparison against other health systems.
By using Singapore as a standard bearer for COVID-19 detection, his team can then estimate how many cases are likely going unnoticed in other countries. Relative to Singapore, other nations with a high capacity for disease surveillance—like the U.S., Japan, Thailand, and afflicted countries in Europe —may only be catching 38 percent of travel-related cases, based on Lipsitch’s model.
That’s due in part to Singapore officials’ dogged efforts toward border screening, but also because the island country is dense. Its 5.6 million people are packed into an area on par with Charlotte, North Carolina, which makes it easier to trace the contacts connected to individual cases.
“If there are [COVID-19] carriers coming into the United States that aren’t detected, which is likely the case, Singapore would probably have an equivalent number,” says Scott Gottlieb, a former U.S. FDA commissioner and resident fellow at the American Enterprise Institute who thinks Lipsitch’s research is on the right track. “But they'd be unmasked sooner in Singapore because it's a smaller, more densely populated nation.” For this reason, Gottlieb says that Singapore offers an omen for major cities in the U.S., rather than the nation as a whole.
“They're a good bellwether for what happens when an outbreak steps into a very sophisticated health care system and advanced economy in a dense environment,” he says.
There are caveats to Lipsitch’s models, including that mild or asymptomatic cases of COVID-19 may not be accounted for. Plus, these models are yet to be published in a peer-reviewed journal, though the process is under way. But their results mirror what others are predicting about the leaky nature of travel restrictions and what they might mean for preparedness in at-risk countries in places like Africa. The model predicts that up to 89 percent of imported cases might be going undetected in countries with a low capacity for disease surveillance.
And undetected cases passing across borders are such a concern because they represent the first step to local transmission, the stage where an outbreak becomes disconnected from importation and begins to perpetuate freely in a region.
On Thursday, researchers in Singapore, including one from the ministry of health, reported eight of the nation’s 84 cases are yet to be linked to any clear exposure. Meanwhile, the COVID-19 tally doubled in South Korea to 204, due to an outbreak at a church in Daegu tied to a 61-year-old woman known only as “Patient 31.” But the source of her infection is currently unknown.
“Although the total number of cases outside China remains relatively small, we’re concerned about the number of cases with no clear epidemiological link,” said Tedros Adhanom Ghebreyesus, director-general of the World Health Organization, said at a Friday press briefing. “Apart from the Diamond Princess cruise ship, the Republic of Korea now has the most cases outside of China. We’re working closely with the government to fully understand the transmission dynamics that led to this increase.”
More than 80 new cases reported Friday are tied to Patient 31, and Korea’s Centers for Disease Control & Prevention described this incident in Korea’s fourth largest city as a super-spreader event.
“There is a real risk,” Lipsitch says, “that there's ongoing transmission in other countries, which if it remains undetected is going to result in more and more cases.”