From COVID-19 to Hantavirus and Ebola: Why Access to Non-Traditional Data Remains a Critical Gap in Outbreak Preparedness

This piece was originally publsiehd on Medium here.

The emergence of Hantavirus cases appeared, at first glance, to be a localized public health incident. Yet, when viewed alongside recurring Ebola outbreaks, growing concerns about avian influenza, and other zoonotic disease threats, these events highlight a broader lesson from COVID-19: our vulnerability to infectious disease outbreaks is shaped not only by the pathogens themselves but also by the preparedness of our data ecosystems and our ability to translate information into timely action. In particular, responsible access to non-traditional data sources -including mobility data, online search behavior, social media activity, transaction records, crowdsourced information, and other digital traces- has become critical in complementing traditional surveillance systems. These data sources can provide earlier, more timely, and more granular insights into emerging risks, helping decision-makers detect outbreaks sooner, understand behavioral dynamics, target interventions more effectively, and strengthen overall preparedness and response efforts

As such, these recent outbreaks provide a useful lens through which to revisit some of the insights of our recent research on non-traditional data and pandemic preparedness. In the below, we share key insights from our analysis of the COVID-19 response, reflect on their continued relevance in the context of the current Hantavirus and Ebola outbreaks, and outline three recommendations to help ensure that the barriers, delays, and missed opportunities of past crises are not repeated again.

This article is not intended to be a comprehensive assessment of either (on-going) outbreak -such an undertaking will require more extensive epidemiological, operational, and governance analysis (which we recommend). Rather, our objective is to highlight a number of emerging warning signs and recurring challenges that deserve more serious attention.

First and Last Mile Challenges

COVID-19 fundamentally transformed the role of non-traditional data in public health. Governments, researchers, and health agencies increasingly relied on new data streams including mobility traces, wastewater monitoring, wearable devices, social media activity, online search behavior, and transaction data. Our research found that 90 percent of epidemic modelers surveyed used at least one non-traditional data source during the pandemic, most commonly mobility, wastewater, and social media data.

Yet despite these innovations, COVID-19 also exposed deep structural weaknesses in the data ecosystem. The problem was not simply that institutions lacked access to data. Rather, even when data was available, institutions often struggled to use it effectively. We have described this as the dual challenge of “first-mile” and “last-mile” failures.

The “first mile” concerns access: identifying, obtaining, and sharing relevant non-traditional data quickly enough to support response efforts. During COVID-19, these barriers were everywhere. Governments struggled to negotiate ad hoc and bilateral agreements with telecom providers. Researchers relied on informal networks to access mobility data. Data formats varied across jurisdictions. Metadata was incomplete. Quality standards were inconsistent.

Equally importantly, COVID-19 also exposed the “last-mile” problem: the inability to transform data into timely and trusted action. Policymakers frequently lacked the institutional mechanisms, technical literacy, or decision frameworks needed to interpret and operationalize insights generated from increasingly complex data streams. This last mile was–and remains–a critical bottleneck in our collective efforts to better unlock public value from the unprecedented amounts of data today being generated and stored.

Hantavirus and Highly Datafied Environments

The Hantavirus outbreak underscores why all these lessons remain urgent and how many of the central challenges remain unaddressed.

Cruise ships are, in many respects, highly datafied environments. They generate enormous quantities of information: passenger movement records, environmental monitoring data, ventilation system analytics, food and sanitation inspections, booking records, and medical logs. In theory, these environments should allow for rapid detection and containment of outbreaks. Yet outbreaks on ships continue to reveal gaps in access, coordination, transparency, and preparedness.

One of the central challenges in outbreaks like Hantavirus is environmental uncertainty. Unlike COVID-19, where respiratory transmission dominated public concern, Hantavirus often involves interactions between humans, rodents, ventilation systems, waste management, and environmental conditions. This demands a much broader data ecosystem than conventional disease surveillance alone, including non-traditional data. With the Andes variant in particular, where human-to-human transmissions play a key role (though to a lesser extent than with COVID-19), we must also add in many of the societal variables that were important for that earlier pandemic.

Amid this complexity, environmental intelligence and monitoring data become essential. So too does behavioral data: where passengers traveled, what facilities they used, how crew movements intersected with operational systems, and how sanitation protocols evolved over time. Crucially, non-traditional data can also shed light on the social behavior of individuals in these confined environments -their contact patterns, gathering habits, dining and recreational interactions, and information-sharing networks- all of which shape transmission dynamics in ways that conventional surveillance alone cannot capture. Increasingly, these forms of information may come not only from traditional inspections but also from sensor networks, logistics systems, occupancy analytics, wastewater analysis, and even indirect indicators such as purchasing patterns or maintenance records.

Having access to these data sources is key. But as the prevalence of “last-mile” challenges indicates, simply possessing such data does not guarantee preparedness. COVID-19 already demonstrated that data without governance produces fragmentation rather than resilience. Many institutions lacked clear stewardship structures to coordinate data access, assess data quality, or align information flows with operational decisions. In many countries, critical insights were trapped in silos between ministries, private companies, hospitals, and research institutions. These are all shortcomings we must avoid this time if we are to prevent the emergence of another pandemic.

Ebola and Social Intelligence

Recurring Ebola outbreaks in the Democratic Republic of Congo illustrate a complementary dimension of this challenge. While the cruise ship represents a datafied, infrastructure-rich environment, the DRC context highlights how non-traditional data matters most when the barriers to containment are fundamentally social.

Experience from successive Ebola responses has shown that community engagement, cultural beliefs around burial practices, and trust in health authorities are decisive factors in outbreak trajectories: without local cooperation, contact tracing, and safe burial, the virus can rapidly spread, while stigma and misinformation remain major barriers to early treatment and containment.

Importantly, a range of innovative platforms and mechanisms have emerged to capture precisely this kind of social and behavioral intelligence. UNICEF’s U-Report platform, for example, has empowered youth and community mobilizers to capture real-time feedback on symptoms, track evolving beliefs, and disseminate preventive health information. Community feedback mechanisms compiled by organizations such as the IFRC have systematically categorized rumors, beliefs, and observations, providing health organizations with actionable insights to adapt their messaging to local contexts. Meanwhile, the Social Science in Humanitarian Action Platform (SSHAP) has analyzed behavioral data to identify specific social bottlenecks during outbreaks. More recently, Flowminder has provided access to monthly population variations and detailed population flows in the Ituri (DRC) province to allow for better understanding of movement patterns and connections between health zones in an area with many international borders, conflicts, population movements, etc.

These examples underscore that non-traditional data is not only a matter of sensors and digital infrastructure; it also encompasses the social intelligence needed to understand how communities perceive, respond to, and engage with public health interventions, insights without which even the most technically sophisticated response will fall short. They also demonstrate that innovation alone is insufficient. Effective preparedness requires both first-mile readiness, the ability to access and share relevant data rapidly, and last-mile readiness -the institutional capacity to translate those insights into timely action. Unfortunately, barriers in governance, coordination, and preparedness delayed both.

Recommendations

A key lesson learned from all of the above is the need to move beyond emergency improvisation toward institutionalized readiness. Preparedness cannot begin only when a crisis emerges. It must be built beforehand through governance frameworks, data-sharing agreements, simulation exercises, interoperable standards, and dedicated stewardship roles.

Our paper made a few specific recommendations that could enable greater preparedness. In particular:

• First, we called for the institutionalization of “data stewards.” These are individuals or teams responsible for building partnerships, conducting risk assessments, facilitating cross-sector coordination, and ensuring that data can move responsibly and effectively during crises. In a future hantavirus-like outbreak, such stewards could help bridge the gap between cruise operators, public health authorities, environmental agencies, laboratories, and researchers.
• Second, and equally important, is the cultivation of what we call “data solidarity.” During COVID-19, many critical data-sharing arrangements depended on goodwill, temporary agreements, or exceptional circumstances. Yet future outbreaks–whether hantavirus, avian influenza, antimicrobial resistance, or other climate-linked or zoonotic diseases– will require more durable frameworks for responsible collaboration between public and private actors.
• Finally, public health systems must also become capable not only of collecting data but also of asking better questions and operationalizing insights from answering those questions rapidly and legitimately. As our paper argues, question literacy — the ability to formulate the right questions and align data with decision-making needs — may be just as important as technical analytics. This is especially critical as the world faces an increasingly diverse set of public health threats, ranging from hantavirus and avian influenza to recurring Ebola outbreaks, antimicrobial resistance, and other climate-linked or zoonotic diseases. Each of these crises requires not only biomedical preparedness but also the ability to rapidly mobilize trusted, interoperable, and responsibly governed data ecosystems across borders and sectors.

New Warning Signals

The Hantavirus and Ebola outbreaks are more than isolated epidemiological events. They are warning signals. They reveal how fragile and uneven our preparedness systems remain even so many years after COVID-19.

This warning is made all the more urgent by the sharp contraction in global health funding now underway. Massive cutbacks in official development assistance from major donor countries, the withdrawal of the United States from the WHO, and the dismantling of USAID have created what the WHO Director-General has called “one of the most difficult years” for the organization. External health aid is projected to drop by 30 to 40 percent compared with 2023, and WHO survey data from 108 low- and middle-income countries indicate that funding cuts have reduced critical services — including disease surveillance and health emergency preparedness — by up to 70 percent in some settings. The consequences are already visible in the current Ebola response: in eastern DRC, the IRC reports that US-funded health and surveillance activities ended in March 2025, forcing a reduction from five health areas to just two in the outbreak’s epicenter and that the surge in cases likely reflects transmission that went undetected due to weakened surveillance. As the National Academy of Medicine has observed, the persistent cycle of panic, neglect, and repeat continues: despite the enormous costs of successive outbreaks, global political will remains insufficient to sustain the investments in preparedness that could prevent them. The erosion of funding for the very institutions and data systems on which outbreak detection depends makes the arguments for data readiness, stewardship, and solidarity not merely academic but existentially urgent.

The next crisis may not resemble COVID-19 at all, and it already might be here. Slower-moving, more localized, environmentally driven, or intertwined with infrastructure and climate systems — as recent concerns around hantavirus and recurring Ebola outbreaks remind us. But the underlying challenges will persist: can societies build the institutional readiness, governance structures, and collaborative data ecosystems needed to transform information into timely, trusted, and equitable action?

COVID-19 gave us a painful preview of both the possibilities and limitations of data-driven public health. The hantavirus outbreak reminds us that those lessons remain all too relevant–and that the work of unlocking the vast potential of data to address social crises likewise remains unfinished.

Daniela Paolotti is a Senior Research Scientist in the Data Science for Social Impact area at the ISI Foundation in Turin, Italy, and Scientific coordinator of Influenzanet, network of Web platforms for flu surveillance in Europe.

Stefaan Verhulst is the Co-Founder of The GovLab, and the DataTank. He is also a Research Professor at the Tandon School of Engineering, New York University