This is a guest post from Ben Bellows, PhD (UC Berkeley, epidemiology), currently a researcher at the Population Council in Washington DC and a co-founder and the Chief Business Officer at Nivi Inc., a digital health company empowering consumers in emerging markets. Nivi is supporting the COVID-19 response here.  

Background: the problem

Disease outbreaks are as much a social phenomenon as a biological one. Rumor, innuendo, and public sentiment drive disease transmission dynamics. Covid-19 is no different; the fact checking website, Snopes.com, has a dedicated “covid-19” tag to run the equivalent of public health containment and mitigation on misinformation.

As the World Health Organization reported 2nd February, this pandemic, similar to past epidemics like SARS in 2003 when the term was coined, risks spawning an “infodemic” that exacerbates disease control and treatment efforts (e.g. there is no high quality evidence that hydroxychloroquine leads to significant improvements in COVID-19 outcomes, drinking hot water will not kill the virus, and this video does not show mass COVID-19 graves in Italy).

Drawing lessons from the 2013-16 West Africa Ebola epidemic, we already see health authorities trying to balance centralized mechanisms to promote consistent and high quality messaging with decentralized programmatic communication that is flexible and adaptable to local needs, as Gillespie and colleagues recommended post-Ebola.

As a part of the global response, on March 20th, the WHO and Facebook launched a WhatsApp chat bot and resource page to improve information quality, allowing anyone with access to WhatsApp to learn more about coronavirus and receive updates. This proactive strengthening of health authority messaging pairs well with effective efforts to weaken transmission of misinformation on social media platforms (e.g. YouTube has a 24-hour incident-response team to remove misinformation and Facebook partnered with the International Fact Checking Network awarding grants to fact check Coronavirus misinformation).

National governments and corporate partners have also launched digital messaging services including in India, Kenya, and South Africa. These messaging services, and similar web platforms the US CDC Coronavirus Self-Checker, convey consistent and vetted health information intended to inform citizens and empower healthy decisions.

Next phase in pandemic response

Anticipating the next step in the COVID-19 response, digital health companies are launching consumer-facing COVID-19 messaging for specific populations. One company that I co-founded in 2016, Nivi, began to develop messaging after noticing an uptick of in-bound questions from its users.

Nivi is an artificial intelligence (AI)-enhanced consumer-facing digital health platform that offers a subscription-based platform to nonprofits, private companies, and public agencies engaging individuals with personalized health conversations via popular digital messaging channels (e.g. Facebook Messenger, WhatsApp). The conversations have an informal tone, may include a fun quiz or emoji responses, and often involve a behavioral nudge or call to action like a geo-located referral to nearby health products or services.

This sustained user engagement maps along a user’s health journey (see Figure 1). Through its subscription-based partnerships and its own efforts, Nivi’s aim is to build broad health awareness by ironically empowering users to explore their individual health interests, generate informed and action-oriented demand for high quality healthcare, and improve health outcomes at scale. As of March 25, Nivi has cumulatively engaged more than 1.6 million people through online ads, radio, and on-the-ground campaigns in Kenya (active since 2017) and India, where it launched in August 2019.

Figure 1: General pathway for a Nivi user as they follow a typical health journey

Nivi’s global response to COVID-19

In March, we began receiving in-bound messages from Nivi users in Kenya asking about coronavirus. Given the rapidly evolving risk to broad segments of populations worldwide, we made the critical decision to accelerate development of a global version of Nivi, beginning with COVID-19 conversations that educate users, encourage preventive measures, and develop insights to inform an effective public health response to local outbreaks and misinformation.  

With this global platform, Nivi and its partners can reach at-risk audiences in any country, nudging healthier behaviors and amplifying the public health response to COVID-19. Using an informal confidence-building tone, COVID-19 conversations include a symptom checker, a description of risk factors, tips on how to prevent transmission, other rapidly updated FAQs, and a myth busting quiz.

These conversations were created drawing from the WHO and US Centers for Disease Control and represent a solid technical synthesis of the current state of knowledge on COVID-19. The chatbot demo is currently accessible on Facebook Messenger and the fully automated conversations will launch the first week of April.

A digital user experience in non-digital spaces

Nivi’s technology helps our users to access health content quickly, and our subscription partners to make data-driven decisions for an effective response to the COVID-19 pandemic and “infodemics” of misinformation.

Through Nivi’s custom onboarding process, subscription-based partners can monitor the real time performance of their community mobilization strategies and Nivi users are able to quickly connect to health resources.

Through Nivi’s language processing module, users can be directed to specific conversations based on their interests and health priorities, and through interactive messaging conversations, users learn about proper disease prevention techniques, provide information on their symptoms, and assess knowledge to address gaps or prompt action.

For our subscription partners, Nivi’s anonymized insights enable them to monitor user engagement performance across specific online (e.g. FB ad word campaigns) and offline (e.g. radio programming) channels.

Partners have access to anonymized geo-location and demographic cluster data, allowing for real-time understanding of potential hot spots (e.g. symptomatic populations) and to prioritize resources for specific localities.

Aggregated conversational data enables partners to monitor population-level knowledge on specific topics, address misinformation, understand behaviors and attitudes that might serve as barriers to proper prevention, and treatment, and assess the possibility of contact with those infected with the virus through digital screening.

Finally, Nivi’s push notifications reach users with up to date information, and target cohorts with specific messaging to inform attitudes and promote prevention measures, treatment, and build trust with the health system.

Nivi is not the only digital health messaging firm operating in or near this space. Viamo, Babylon, Ada Health, and Turn.io also aim to engage users in digital conversations that lead to better health outcomes via informed self-care or more effective engagement with the health system.

There are limitations to messaging, beyond the rate limiting fact that Facebook users (as monthly active users) represent only 33% of the planet. Lack of telecoms infrastructure, limited phone ownership, digital illiteracy, unaffordable airtime or equipment, language barriers, psychosocial and gender dynamics are all factors that can constrain the promise of digital technology to quiet the misinformation it also foments. Despite these limitations, we think platforms like Nivi are important in combatting the wave of misinformation that has surfaced and help citizens around the world make good decisions with their personal health and our collective well-being.