Disease outbreaks in the age of Facebook

In the spring of 2014, Vietnam’s state-controlled news media reported that dozens of children had died after turning up at hospitals in the capital, Hanoi, with rashes and high fevers. Doctors said the cause was a measles outbreak – the worst in Vietnamese history.

Parents of victims rushed onto Facebook to share their grief and outrage. So did their friends and neighbours, who wanted to know how the outbreak was spreading, and whether it was even safe to take children to hospital. The authorities distributed information about the measles deaths via leaflets, loudspeaker bulletins and updates on the health ministry’s website, recalls a nurse at a state-run hospital in Hanoi, who requested anonymity in order to speak frankly. “But there was so much information online that some people became panicked.”

Vietnam is a one-party state that has responded efficiently to many natural disasters and health emergencies. During a 2005 outbreak of avian flu, for example, the Vietnamese agriculture minister famously ate chicken on national television in an effort to calm public fears about widespread poultry infections. But when the 2014 measles scandal broke, millions of Vietnamese were using Facebook, compared to almost none just a few years earlier. And the government, which sporadically blocks the network, was unable to extinguish a Facebook-fuelled bush-fire of fear, anger and recrimination that swept the nation.

Much of the criticism landed at the feet of the health minister, Nguyen Thi Kim Tien, who refused to call the deaths an ‘outbreak’ even as the child death toll approached 130. Even before the measles crisis the health ministry had been on the defensive, following a string of previous scandals at public hospitals that had included mistaken diagnoses, faulty vaccinations and cases of bribery in surgery wards. So when Tien said that only 25 deaths were directly attributable to measles – while doctors at a Hanoi hospital said scores had died from measles-related complications – many ordinary Vietnamese viewed her comments as brazen obfuscation.

“The public saw a lot of people dying,” recalls Dinh Duc Hoang, a contributing editor at VnExpress, a Vietnamese online newspaper. Ministry officials, he says, did publicly explain their epidemiological rationale for not declaring an outbreak, “but not enough, especially on social media,” he says. “People didn’t understand and thought the ministry was trying to hide the outbreak.”

Since the turn of this century, the rise of online search engines and social media applications has allowed ordinary people around the world to find information during public health scares at an ever-faster pace. The change has been especially dramatic in the Asia-Pacific region, a global epicentre of internet users, smartphone sales and infectious-disease outbreaks.

This new environment presents a complex mix of opportunities and challenges for health officials. On one hand, increased public engagement during a health crisis can allow officials to communicate more directly with citizens. But every new online platform is also a conduit for spreading criticism or misinformation. The rise of social media makes it “harder for governments to shut down the flow of information, but the information itself may be unreliable,” says Crawford Kilian, a Vancouver-based writer who covers the politics of public health.

Health officials are increasingly aware that online rumours and criticism during a health scare can fuel panic, undercut their credibility and weaken public confidence in the very healthcare systems that are meant to fight disease. And some of them are trying to close the digital gap by engaging their citizens – and critics – on social media. During the measles crisis in Vietnam, for example, senior health officials asked Hoang and around eight other well-known journalists to help improve the health ministry’s social media presence and disseminate official updates to the public during health emergencies.

“Because of my job, I’m famous, and I know a lot of celebrities,” Hoang explains. “I can contact people to get them to post correct information for people to read.”

The premise – helping the ministry distribute ‘correct’ information through social media – was simple enough. But could, or should, a group of journalists and celebrities really help a one-party government restore public faith in a beleaguered health bureaucracy? And why did the ministry need their help in the first place?

SARS and the internet prank

Social media is only the latest digital ‘disrupter’ to change the public conversation around health scares; the first was the internet itself. And the first major health scare in which the internet played a decisive role in shaping public opinion was the severe acute respiratory syndrome (SARS) outbreak – the first influenza pandemic of the 21st century.

SARS appeared in southern China in 2002 and created an international crisis in February 2003, after a 64-year-old medical professor from the Chinese city of Guangzhou checked into a Hong Kong hospital with respiratory failure. A guest staying at the same hotel as the professor then unwittingly infected more than 100 medical and nursing staff at a second Hong Kong hospital, and six other guests unknowingly spread SARS to Vietnam, Singapore and Canada. This was just the first phase of an outbreak that would spread to more than two dozen countries, infect 8,098 people and kill 774.

Within months of the first SARS infection, hundreds of thousands of ‘SARS’ searches had popped up on a newish platform called Google. Baidu, a popular Chinese search engine, also reported during the epidemic that ‘SARS’ was its top search item. Both sites were allowing ordinary people in China and beyond to instantly find information about the disease, rather than wait for updates from a local newspaper or television anchor.

“Back then, it was Web 1.0, where people go to a site and download information, and there’s no sort of participation,” says Randall N Hyer, a risk-communications specialist who worked on a World Health Organization (WHO) risk-response team during the SARS crisis.

But not all the information that turned up in the search engines was accurate, and false rumours quickly went viral – faster, in fact, than the virus itself. One of the most damaging came from an unlikely source: a 14-year-old Hong Kong boy, who announced online that the city would soon be placed under quarantine and designated an “infected port”. His prank post gained attention and credibility because it mimicked a local newspaper’s website. It sparked a frenzy of panicked grocery shopping, which fed perceptions that Hong Kong was under siege.

According to At the Epicentre, a 2004 book coedited by Christine Loh (who is now Hong Kong’s undersecretary for the environment), the mood in the city felt “eerie”, in part because there was widespread confusion around how the disease was spreading. Even a public disavowal of the teenager’s prank by Hong Kong’s then health director, the book states, did little to quell the eeriness.

But the internet wasn’t an entirely bad influence during SARS. For example, as infections piled up in Hong Kong, a group of concerned residents created a crowd-sourced website, sosick.org, showing a list of buildings with suspected or diagnosed SARS cases. Two weeks after the site appeared, the Hong Kong government began publishing the same information on its official website.

“The internet proved to be an invaluable tool for concerned citizens to use in challenging government secrecy,” Loh and her coauthors wrote.

Gabriel Leung, Hong Kong’s former undersecretary for food and health, says the SARS experience helped convince Hong Kong authorities of the need to be transparent with the public about the knowns and unknowns of an outbreak as it unfolds – well before a website such as sosick.org might challenge the government’s credibility. He applied the lesson during the city’s response to its next major public health emergency: an outbreak of H1N1, a deadly respiratory virus known as swine flu.

Swine flu was discovered in April 2009 in the USA, and at the time outbreaks had not been known to occur among humans. But as H1N1 quickly spread to Canada and Mexico, WHO declared it a Public Health Emergency of International Concern, and health officials on other continents braced themselves for its arrival.

It landed in Hong Kong in May 2009, when a 25-year-old Mexican tourist fell ill with swine flu in his hotel. Within hours, health officials had quarantined all of the hotel’s guests and deployed Leung to television stations to explain the rationale for a decision that was certain to make Hong Kong residents deeply anxious.

Internet-fuelled rumours about the disease proliferated, but Leung says he mostly stayed ahead of them by explaining the knowns and unknowns of the disease at daily press briefings, and by promising to show up every afternoon at the same time for as long as reporters had questions.

“So long as you keep your promise, people will not be susceptible, or as susceptible, to the rumour mill, because they know the maximum time they have to wait is 24 hours before you will come out and give the facts again,” he says.

More social media, more anxiety

A daily press conference might have kept rumours and medical misinformation at bay during a health scare in the mid-2000s. But today, when fresh social media trends go viral every few minutes, a 24-hour information cycle can feel like an eternity. Social media can also challenge any official narrative that a health ministry tries to promote.

“Rumour mills will always be there… and social media by and large fuels that,” says Leung, now the dean of medicine at the University of Hong Kong’s School of Public Health. “It’s not all negative, but social media is fast, it’s furious.”

Swine flu, which killed between 150,000 and 575,000 people, was – according to a 2010 study in the journal PLOS ONE – the first global pandemic to occur in the age of Web 2.0, an era defined by ‘participatory’ web and social media. Since then, the dominance of social media has only increased, which has intensified the need for health officials to develop rapid online responses during health scares.

The challenge is now especially stark in countries, like Vietnam, where millions of people are flocking to Facebook or Twitter for the first time but health authorities are only just developing an online presence.

As late as the early 2000s, most Vietnamese found information about public health issues through state-controlled newspapers, television channels or a national network of public loudspeakers that broadcasts state-curated news in most towns and villages at daily intervals. All of those channels still exist, and the government still promotes them aggressively. The difference is that many urban Vietnamese now prefer to find information about health issues on health blogs or Facebook threads – and that some now hardly consult official news sources at all.

The new environment can feel liberating, but also scary, for Vietnamese who go online seeking authoritative information during a health crisis. “Vietnam, 10 years ago, was very poor,” says Nguyen Dieu Huong, an academic administrator in Hanoi and a member of Vietnam’s rising middle class. “We didn’t have the internet or ways to update people about, for example, epidemics. So people didn’t really feel anxious or afraid of epidemics, like they do now.”

During the 2014 measles outbreak, the anxiety bubbled up in a Facebook campaign called ‘Health minister, resign!’ that earned more than 100,000 ‘likes’ – a rare outpouring of public dissatisfaction in a one-party state that imprisons political dissidents. Thousands more Vietnamese logged on to discussion threads on Facebook and Webtretho, a popular web forum for parents. Among other questions, they debated whether the health benefits of bringing children to paediatric hospitals for vaccinations outweighed the apparent risks.

“We were scared to vaccinate our kids, and, at the same time, to not vaccinate them. So we were scared of everything,” says Thai Lan Anh, a Hanoi yoga teacher who participated in those discussion threads.

Lan Anh says her first child, who was born in 2011, had already been fully vaccinated by the time the measles crisis hit. But she decided not to vaccinate her second child, who was born in 2014, out of an abundance of caution – and still hasn’t.

“I don’t think the Ministry of Health has been totally honest [about the risks of diseases and vaccinations],” she tells me between sips of fruit juice at a Hanoi café. “I don’t believe what they say.”

A health minister's image management

Even before the measles outbreak, health minister Nguyen Thi Kim Tien was among Vietnam’s most reviled public servants. An easy way to tell was the extensive focus on healthcare in Meeting Each Other at Year’s End, a hugely popular television satire that is performed once a year on Vietnamese state television. A 2013 skit, for example, showed a patient going to a Vietnamese hospital for a kidney operation, only to find that a doctor had removed the wrong kidney. A year later, a skit showed a health minister – widely assumed to be Tien – being placed on a machine that tested her ethical integrity. She failed miserably.

Against that backdrop, the health ministry quietly reached out to Hoang and other journalists in 2014 to solicit their ideas on improving the minister’s image and social media presence, according to Hoang. (A ministry spokesman, when reached for comment by telephone, asked for questions by email but did not answer them.)

Hoang, whose mother is a doctor, says he felt a duty to help.

“Ms Tien wanted to build an image of the doctors and herself as friendly and close to the public,” he says of his early exchanges with the ministry, which he insisted were informal and entirely pro bono.

Tien was setting up a Facebook page, and she asked if the journalists would mind ghost-writing some posts for her, according to Hoang. The journalists refused because they were too busy, he says. But, as a false rumour about Ebola spreading in Vietnam circulated on Facebook, they agreed to help staunch the flow of misinformation by using their Facebook pages to publicise the ministry’s Ebola updates.

Hoang says he also personally helped the minister improve her online image. First, he advised her to post a more friendly looking photo on her Facebook page. He also recommended that she interact more with ordinary people.

He saw an opportunity when a Vietnamese cancer patient wrote to a local newspaper to complain that a state-funded insurance programme was expiring, thus threatening to deprive her and other cancer patients of live-saving healthcare benefits. Hoang forwarded the patient’s letter to the minister, who drafted a reply that empathised with the patient’s predicament and explained the ministry’s insurance policy in accessible language. Hoang then publicised the exchange on the VnExpress website and his Facebook account, which has more than 30,000 followers.

But even then, the exact effect of the intervention was unclear.

“I made the two letters a campaign for the awareness of cancer and the image of the minister,” he says. “The campaign helped people understand the policy, but I can’t tell if it helped her image or not.”

Digital epidemiology

This problem – of how to use social media effectively to benefit public health – spreads beyond Vietnam.

Careful study of health-related social media trends could eventually help scientists and health authorities better understand person-to-person disease transmission, in part because social media tends to be “highly contextual and increasingly hyperlocal”, says Marcel Salathé, a researcher at the École Polytechnique Fédérale in Lausanne, Switzerland, and a specialist in an emerging field scientists call ‘digital epidemiology’. But for now, he says, they are still struggling to understand whether social media chatter about health problems actually mirrors epidemiological trends or not.

“There have been some successes, but many failures,” says Isaac Chun-Hai Fung, a digital epidemiologist at Georgia Southern University in the USA.

An early, and humbling, experiment in digital epidemiology occurred at Google, where engineers launched the disease-forecasting tool Google Flu Trends (GFT) in 2008. The company planned to analyse Google search data for mentions of symptoms and other telltale words. It would then, it was hoped, crunch the data to accurately ‘nowcast’ the likely contours of flu and dengue outbreaks, two weeks earlier than the Centers for Disease Control and Prevention (CDC) – the gold standard of epidemiological research – could.

That never happened. In 2009, when panic spread across the globe during the swine flu outbreak, GFT underestimated the outbreak’s impact. One potential cause of the error, according to a study in the academic journal PLOS ONE, was that the pandemic occurred in summer rather than winter, the traditional flu season.

Google shrugged off the mistake as a correctable glitch. But during a flu outbreak in December 2012, GFT’s estimate of the percentage of Americans who had influenza-like illnesses was over 10 per cent – a prediction that the scientific journal Nature said had “drastically overestimated” the outbreak compared to CDC surveillance data, which put the actual figure at around 6 per cent. Experts concluded that internet search algorithms were not yet a reliable replacement for the CDC’s traditional epidemiological surveillance. GFT was quietly shelved in 2014, and Google said in a note to the public that it was “still early days” for digital applications that track the spread of infectious diseases.

Other projects have picked up where GFT left off, with some success. Models built using Google search data from between 2005 and 2010 in Thailand were able to “adequately” estimate malaria trends in the country, for example, according to a 2013 study in the Malaria Journal. Another study that year noted that an algorithm built to track health trends on Baidu, China’s primary search engine, successfully predicted the contours of seasonal influenza infections over an eight-month period.

But “social media is still a relatively recent phenomenon in human history,” Fung says. “There are still a lot of unknowns here, both in terms of whether we can invent an algorithm to make better estimates or understand what people really are communicating.”

A meter of credibility

Even so, health officials are still attempting to engage social media chatter in a way that keeps citizens informed about health scares in real time and tamps down misinformation, all without appearing to silence critics in a heavy-handed way or losing the public’s trust. That balance can be hard to strike, even in a rich country with a high rate of smartphone penetration and the highest internet speeds on earth.

That was clear in the spring of 2015, when the South Korean capital, Seoul, became ground zero for the largest outbreak outside the Middle East of Middle Eastern respiratory syndrome, or MERS.

MERS, which was discovered in 2012 in Saudi Arabia, was bound to cause anxiety wherever it landed because it kills 30 to 40 per cent of its victims. Moreover, because the disease’s inner workings are still poorly understood, it can be hard to detect if a doctor isn’t looking for it.

South Korea’s first patient was not diagnosed until he showed up at Samsung Medical Center in Seoul with MERS symptoms – nine days after first showing signs of infection, and after visiting a few other hospitals. That was bad enough. But for more than two weeks after South Korea’s initial MERS case was diagnosed, the government declined to publish the names of MERS-affected hospitals – just as the Hong Kong government had declined to publish basic information about SARS infections in 2003. And as health officials stalled, the outbreak spread to other hospitals.

Korean social media erupted with criticism, and on 4 June, a South Korean journalist published the list of affected hospitals on Pressian.com, an online news portal. The government, now in full damage-control mode, quickly followed suit, and a Samsung official later apologised for the hospital’s handling of the outbreak.

But the government’s and the hospital’s credibility was already tarnished. It didn’t help that, on the same day Pressian.com published the hospital list, South Korea’s Centers for Disease Control and Prevention made its Twitter account private, prompting a backlash from the scientific community and fuelling speculation that the government had something to hide – even if it didn’t.

Poor outbreak response

I visited Seoul during the crisis, and the city did not feel panicked. Perhaps the biggest visual reminders of the MERS virus, at least for non-native speakers, were the public health posters with images of a camel – the animals were the origin of the disease – that I saw in Seoul’s Incheon Airport. The camel didn’t look very scary. Yet on the streets and in the metros, many people were wearing face masks, reflecting the widespread public unease about the virus that was bubbling up on social media.

Over the course of a week, I asked South Korean journalists about the MERS situation, and many of them were highly critical of their government’s response to the outbreak. Officials should have released the names of affected hospitals more quickly, they said.

Lee Won-jae, a high school student, had a similar complaint. “Time has passed” since the first MERS infection, Lee told me, standing a few blocks from Samsung Medical Center in suburban Seoul on 12 June, at the height of the government’s MERS-news blackout. “Things should be wrapped up, but everything is still ongoing.”

At the time, President Park Geun-hye was still recovering from handling a ferry accident that had happened in April 2014, in which 304 South Koreans – mostly teenage students – were killed. The government had faced criticism over its rescue efforts. Because of the MERS scandal, Lee said, “there’s a bigger distrust of the government now than before”.

A few months later, as South Korea’s MERS epidemic was tapering off, a team of researchers from Hong Kong and the USA published a briefing about it in the Western Pacific Surveillance and Response Journal. The South Korean government’s response to MERS, they concluded, was “reminiscent” of the Chinese government’s response to SARS in 2003.

Gabriel Leung, the former Hong Kong official, agrees. He says one lesson may be that having a prosperous economy does not necessarily prepare health officials to respond smoothly to a 21st-century outbreak in real time and in a way that doesn’t damage the government’s credibility.

“Despite the high-income nature of the setting, despite excellent medical care, there are some lessons that societies generally need to learn experientially rather than just from a theoretical perspective,” he says of the MERS crisis.

Lessons learnt

In Vietnam, another public health scare bubbled up this spring, when WHO reported that an Australian tourist had been diagnosed with Zika virus after vacationing in the country. But as Facebook rumours began to swirl, the health ministry’s information chief, Nguyen Dinh Anh, began posting international news stories on his Facebook page and explaining what Vietnam was doing to prepare for a possible outbreak.

Hoang, of VnExpress, says the health minister’s Facebook page is still relatively sparse, and should be updated and managed by a professional. But the information chief’s response to Zika, he adds, seems to show that the ministry has at least recognised the importance of being transparent about public health risks and engaging the Vietnamese public on social media.

This time around, “they’re posting on their own, without our help,” Hoang tells me between drags of a cigarette in a Hanoi café. “They won’t have the type of crisis they had with the measles outbreak.”

Ultimately, he adds, it would be difficult – if not impossible – for the health ministry to win the public’s complete trust, for the simple reason that Vietnamese bureaucracy is notoriously corrupt. But he says the ministry could perhaps work to become a more reliable conduit of information during future health crises.

“The people still need it, and that’s the only source.”

This article first appeared on Mosaic.