Individually Customized Digital Tool to Counter Vaccine Hesitancy

A digital game that leverages adaptive learning profiles to combat misinformation and increase COVID-19 vaccine adoption globally.

Farhaan Ladhani, Co-Founder and CEO of Digital Public Square

Widespread vaccine hesitancy is a complex global phenomenon that threatens to undermine efforts to gain control over the COVID-19 pandemic. Whereas 70-90% of the world’s population will need to be vaccinated to achieve herd immunity, it is estimated that up to a quarter of Americans do not plan to be vaccinated, and in some countries population-level vaccine acceptance is as low as 15%. A key contributor to this issue is rampant online misinformation campaigns, against which the susceptibility of an individual is affected by a range of psychological, socio-economic, and environmental factors.

Importantly, when these factors are not addressed, misinformation interventions typically fail to have an effect and have even backfired in some instances. Treatment of vaccine hesitancy as a problem of information deficiency alone overlooks key influential factors including user values, affective and motivational reactions, along with contextual barriers stemming from historical mistrust of medical institutions in marginalized communities. Globally this problem is pervasive amongst healthcare providers in formerly colonized settings; for instance, a recent study in Nigeria found vaccine acceptance at 28% amongst healthcare workers. Overcoming vaccine hesitancy is therefore even more critical as the world focuses on ending both the COVID-19 pandemic and decolonizing public health.

Our target audience are individuals that are vulnerable to misinformation about COVID-19 and/or are vaccine-hesitant. Unfortunately, these populations tend to have the greatest incidence rates of COVID-19, and across the USA and Canada are disproportionately represented in communities of colour and underserved communities. Globally, our target audience extends beyond the general public, with vaccine hesitancy prevalent amongst healthcare providers and complicated in settings where governments themselves may be perpetuating false narratives. 

Our baseline work included developing a vulnerability assessment to identify themes and factors affecting an individual’s vulnerability to misinformation, consulting groups on different content design elements, and then finally piloting the app with these communities to assess whether we were achieving robust learning effects across diverse groups. We plan to take a similar approach with new communities/regions in this next phase, and as we continue to scale, deployment of our vulnerability assessments will help us strategically define our target users for every new context we enter. 

Altogether, not only does this solution aim to improve vaccine acceptance rates to collectively move past this pandemic but is rooted in equity and prioritization of vulnerable populations everywhere. 

Our solution is in the form of a digital platform that is publicly available, open access and free to use worldwide. The knowledge provided by the app, curated in partnership with leading infectious disease and epidemiological experts is also a public good that is openly accessible to all.

Our goal is to improve health literacy of our end users, which we anticipate will translate into lower rates of vaccine hesitancy and curb the spread of misinformation online. Evidence to date suggests that our intervention is effective at achieving these outcomes, as users have demonstrated positive affect, increased knowledge and increased intent to receive a COVID-19 vaccine irrespective of ideological background or baseline knowledge. 

Specifically, our target audience are individuals that are vulnerable to misinformation about COVID-19 and/or are vaccine-hesitant. Unfortunately, these populations tend to have the greatest incidence rates of COVID-19, and across the USA and Canada are disproportionately represented in communities of colour and underserved communities. Globally, our target audience extends beyond the general public, with vaccine hesitancy prevalent amongst healthcare providers and complicated in settings where governments themselves may be perpetuating false narratives. Critically, our findings to date hold for vulnerable populations of interest, which in our Canadian pilot study included, among others, young participants aged 15-24, and those who self identified as Metis, Inuit, or First Nations.

The platform has been tested vigorously using RCT designs in both Canadian and US contexts, and in its current iteration we believe our product to be robust, adaptable, and scalable to new contexts. We are now poised to start scaling up our platform into more communities around the globe and are currently exploring options to do so with partners in the United States, South Africa, Myanmar, Ethiopia, Pakistan, Tanzania, India, and Brazil. 

At this stage, we are looking for partners who have connections with communities at the grassroots level, and whose networks can be leveraged to deploy our program at scale. This includes community health worker programs, faith organizations and other digital health initiatives which have the potential of accelerating the reach of our app to our target communities.

As we continue to scale, we will also work with partners to expand the content on the platform to address other health information challenges, including vaccine hesitancy more broadly.

Our key indicators for success are reach (the # of people engaging on the platform), knowledge score, emotional affect, and stated intent to receive a COVID-19 vaccine. To date, over 180,000 Canadians have engaged with our tools, and we are observing both positive affect and significant increases in knowledge score in participants across the ideological spectrum and from marginalized communities. For example, 86.7% of people who were given a repeat question in our pilot of the tool, got the answer correct on their second attempt, a finding that held true after controlling for participants’ baseline knowledge and across the ideological spectrum. Finally, repeated measures analysis showed that intent to receive a COVID-19 vaccine significantly increased in the game condition demonstrating that the DPS framework is an effective tool for increasing intended vaccine adoption.

Our key distribution challenge will be to build partnerships where we can leverage trusted channels to reach our target end users to increase community buy-in. Specifically, we are looking for partners who have connections with communities at the grassroots level, and whose networks can be leveraged to deploy our program. This includes community health programs and other digital health initiatives working with community leaders, which have the potential of accelerating the effectiveness of our app with target communities. We are mitigating the distribution component of this challenge through highly effective paid distribution and seeing the evidence of current network effects from the product (+10% actual organic sharing by product users). Subsequent barriers will be advertising costs, which can be defrayed through strategic partnerships with social platforms. We have successfully negotiated some of these already, and scaling investments from this initiative will support that work.  From a technical perspective, our ability to leverage adaptive models from one context, to the other, is expected to be a challenge. We are mitigating to some extent with the collection of context-specific training data. Resources from this program will support the solution of expected technical hurdles related to operationalizing these models.

• University of South Carolina – epidemiological specialists; Platform content contributors
• Alberta Machine Intelligence Institute: AI/ML experts that support development and deployment of adaptive learning profiles

Our key distribution challenge will be to build partnerships where we can leverage trusted channels to reach our target end users to increase community buy-in. We believe that the MIT Trinity Challenge platform will not only provide us the resources to fund the scale of our program and required  partnership building, but will allow us to strategically connect with other like minded institutions that are tackling similar challenges. We believe that our platform and organization bring a number of strategic advantages to the table including a robust platform and implementation model, and strength in data analytics and machine learning. However we need the right partners to help us optimize the adoption of this program in localized contexts. Organizations that are countering misinformation at the grassroots level, such as community health worker groups, faith-based organizations, and other related health campaigns will be critical to getting our intervention to end users in the most vulnerable communities.

We are looking for partners to: 

1. Help us optimize the adoption of this program in localized contexts. Organizations that are countering misinformation at the grassroots level, such as community health worker groups,  faith-based organizations, and related health campaigns will be critical to getting our intervention to end users in the most vulnerable communities. Accordingly we have interest in pursuing partnerships with organizations that have deep country-level networks including CHAI, PATH and the Gates Foundation. 

2. Participate in a cross-exchange of learnings and expand our digital networks to communities of interest. This includes organizations that have similar values and strengths in the digital analytics space, including Facebook, Google, the Patrick J McGovern Foundation and Zenysis.