First Draft Case Study: Understanding the Impact of Polio Vaccine Disinformation in Pakistan

"It's one thing to read an article about vaccines potentially making children fall ill, and another thing to see someone passionately warning people against vaccinating their own children. Social media has a unique ability to spread emotionally-charged messages across cities, countries and platforms." - Sarika Bhattacharjee

On April 22 2019, several staged videos went viral on Twitter and Facebook falsely claiming that children had become gravely ill after their polio vaccinations in the province of Khyber Pakhtunkhwa (KP) in northwest Pakistan. This case study uncovers how the videos spread on social media, the reasons behind the history of vaccine hesitancy in Pakistan, and how the footage became an engine for vaccine misinformation and mistrust, derailing efforts to immunise millions of children.

As is explained here, KP's struggle with vaccine acceptance can be linked to a fake Hepatitis B vaccination campaign that the United States (US)' Central Intelligence Agency (CIA) orchestrated in 2011 to obtain Osama bin Laden's family DNA. As a result, many Pakistanis in the region conflate polio vaccination with illegitimate and dangerous Western influence in the country's affairs. In addition, false rumours that the polio vaccine contains ingredients that are forbidden in Islam (e.g., pork) or that induce sterilisation or make children sick have circulated for several years in the country.

It was against this backdrop of vaccine hesitancy that the April 2019 viral video incident occurred. While most of the rumour-ridden videos have since been taken down, reshares received 24,444 Twitter interactions (retweets and likes) within the first 24 hours of their initial posting. Several verified Twitter accounts also reshared the clips. The official account of the prime minister's political party pointed out the fraudulent nature of one of the clips but shared it anyway, where it has since received nearly 50,000 views. Furthermore, in the immediate 24 hours after the videos were created, public Facebook posts that shared the fake videos (either pushing the narrative that the vaccine was dangerous or debunking the claim) reached a total of 990,493 interactions, between likes and shares.

In the hours that followed, in the city of Peshawar, KP, alone, the parents of 25,000 children brought them to the hospital, fearing they had been poisoned. By the end of the week, the number had reached 45,000. In the same week, a mob of 500 people set fire to a health clinic in Peshawar, and mosque loudspeakers broadcast rumours that polio medicines were "poisonous". Two police officers and one female health worker died in the maylay. The provincial health department called the disinformation campaign a "coordinated conspiracy to disrupt the polio immunisation drive", and police arrested at least 12 people over the videos.

The case study describes the impact the misinformation incident had on polio vaccine uptake. In Nowshera, KP, vaccine refusals increased from 256 instances in March to 88,000 in April. According to Pakistan's National Emergency Operations Center dashboard, there were 990,000 vaccine refusals in all of Pakistan in the month of April. The country had been on course to successfully immunise 39 million children under the age of 5 before the staged videos went viral. But Pakistani authorities suspended their anti-polio campaign on April 27, 5 days after the misinformation outbreak. Since then, 2 million children have gone unvaccinated. The impact has been severe: 144 wild polio virus (WPV) cases were reported in Pakistan in 2019, compared to a total of 12 cases in all of 2018.

On July 9 2019, the prime minister's former focal person for polio eradication announced that an emergency vaccination campaign would be held from July 15 to July 18 in 12 districts in which polio cases had been discovered. This came after 4 new cases emerged in Balochistan and KP, all of which were attributed to parents' refusals to vaccinate their children. Yet a July 21 2019 article described vaccine-hesitant Pakistani parents use of the same pens used by health workers to indicate on a fingernail that a child has been vaccinated to avoid actually getting their children vaccinated.

To understand how this all happened, the case study presents Claire Wardle's Trumpet of Amplification, which is visualisation for understanding how misinformation is amplified as it spreads from closed channels of communication (e.g., online groups like Whatsapp), to open networks and professional news media. The authors then explain their methodology, which involved collecting data on engagements and interactions from January to June 2019 (that is, posts that contained keywords and hashtags related to polio). They used Crowdtangle to collect Facebook, Instagram, and Reddit posts, while the Python programming language was used to automatically collect tweets from the Twitter Search API. They focused on calculating keyword frequency spikes over time and analysing top retweeters and posts that gained high interactions.

The historical data from Twitter and Facebook informs an interactive timeline of how the disinformation went viral and who amplified the rumours. (Also see Figure 1, which contains Google Trends data showing the relative popularity of particular search terms between April 19 and April 25 2019. It reflects a notable spike in the relative frequency of searches for terms related to polio and vaccines in Pakistan on April 22.) Based on the data, the authors do not believe that hashtags played a major role in amplifying the rumours. Rather, on Twitter, the sources amplifying the rumour appear to be Pakistani news organisations and pro-vaccine figures who gave an audience to the rumour by sharing articles and videos about it.

From this analysis, the authors cull out key themes in messaging that emerged across all platforms: a focus on the vaccines' effects on children, anti-Western and "War on Islam" sentiments, a lack of faith in government systems, and gunpoint vaccinations. (An explanation of the latter theme: Many social media posts included images of polio vaccinators being assisted by armed police officers there to ensure the vaccinators' safety; the posts inaccurately conflate the weapons in the photograph with the idea that polio vaccinators are forcing parents to vaccinate their children.)

The case study concludes that this incident can serve as a learning opportunity for social media platforms, health authorities, and the news media. Recommendations include:

• After an outbreak of misinformation, social media platforms should release the archive of posts and accounts that have been blocked, demoted, or removed to health authorities, researchers, and journalists to help them understand how to stem the growth of rumours.
• The case study found that much of the amplification of the rumours that occurred was due to professional media outlets sharing videos with misleading or underdeveloped captions and reporting. This can be avoided through training seminars in best practices for social media content management: verification and responsible reporting.
• Visual misinformation should be monitored more closely, perhaps using the artificial intelligence (AI) capabilities of Google's DeepMind.
• Twitter, Facebook, and YouTube should work together to develop a cohesive strategy that enables them to take down the same sources of misinformation as quickly as possible.