Public search behavior during the Bundibugyo outbreak exposed a second emergency
The 2026 Bundibugyo virus disease outbreak in the Democratic Republic of the Congo and Uganda unfolded as both a virologic emergency and an information emergency. Response teams faced conspiracy-driven violence, treatment centres were destroyed, and public uncertainty intensified around a virus for which no licensed vaccine exists. In that setting, the questions people ask online are not a curiosity. They are a real-time record of fear, perceived risk, and the gaps left by official communication.
This analysis mapped public information demand across six Ebola outbreaks from 2014 through 2026 using Google Autocomplete and Google Trends. The design captured 7,347 unique autocomplete queries derived from 83 seed terms in English and French, then tracked how search interest shifted over time. The result was a consistent picture of outbreak-driven curiosity centered on danger, transmission, and the meaning of the disease itself.
The pattern is clinically relevant because information demand often moves faster than institutional messaging.
Query patterns centered on risk, transmission, and what people feared most
Autocomplete suggestions clustered around the questions that matter most when a hemorrhagic fever begins to spread: How dangerous is it, how does it spread, and what should people believe? The dominance of risk quantification and transmission concerns suggests that users were not searching for abstract background material. They were searching for immediate, personal answers about exposure and survival.
Across outbreak periods, Google Trends showed moderate to strong co-movement between misinformation-related search terms and overall Ebola search interest, with Spearman ρ values ranging from 0.359 to 0.697 and nominal p values below 0.001. In practical terms, interest in Ebola rose alongside interest in misleading or conspiratorial content rather than separate from it. When a virus becomes the subject of rumor, search behavior can amplify confusion at the same pace it amplifies awareness.
For outbreak control, this matters because the same attention that can direct people toward accurate guidance can also pull them toward false narratives. A public health response that only counts cases and contacts but ignores search behavior misses a major channel through which fear spreads.
French-language searches carried a heavier burden of fear and misinformation
The clearest divergence in this analysis appeared between English-language and French-language suggestion sets. French-language queries contained six times more misinformation and conspiracy content than English-language queries, 4.9% versus 0.8%. Fear-related content was nearly nine times more common in French-language suggestions, 3.5% versus 0.4%.
Those differences are not cosmetic. They point to a communication gap shaped by language, trust, local context, and the channels through which rumor travels. In the DRC and neighboring regions, French-language information ecosystems often serve populations that are also navigating conflict, displacement, and fragile health infrastructure. When fear dominates search behavior in one language set more than another, it suggests that outbreak communication is not reaching audiences evenly or with the same credibility.
The language effect also argues against one-size-fits-all messaging. Infodemic management designed for English-speaking audiences may fail where the public is searching in French, where fear and conspiracy language appear far more frequently. Public health teams need communication tools that are linguistically adapted, culturally grounded, and fast enough to compete with rumor in the first days of an outbreak.
Why search data should influence outbreak response planning
Google Autocomplete and Google Trends do not replace field surveillance, virology, or contact tracing. They do something different and valuable: they reveal what the public wants to know while the outbreak is still unfolding. That makes them useful for communication planning, especially when violence, distrust, and misinformation threaten response operations.
The outbreak described here involved a species lacking a licensed vaccine, which raises the stakes of public cooperation even higher. When no vaccine is available, response teams depend heavily on isolation, case finding, safe care, and the public’s willingness to accept those measures. Fear-driven searches can erode that cooperation if they are not countered with clear, local, and timely explanations of transmission risk and protective behavior.
The co-movement between misinformation searches and Ebola interest also suggests that rumor tracking should be treated as part of epidemic intelligence. Search volumes can flag when a population is moving toward panic, disbelief, or hostile interpretation of response efforts. In an outbreak where treatment centres may be attacked and health workers viewed with suspicion, that intelligence has direct operational value.
Communication strategies need to match the language of risk
This dual infodemiology analysis makes a narrow but consequential point: information demand is not uniform across languages, and the French-language ecosystem carried a much heavier load of fear and conspiracy content. Communication teams cannot assume that translated material alone will close that gap. The content, tone, and delivery route of messaging need to match the questions people are actually asking in each setting.
Risk quantification and transmission concerns dominated the search landscape, which means public-facing messages should answer those questions plainly and early. How infectious is Bundibugyo virus? What exposures matter? What symptoms require action? Where should people go, and what should they avoid? The search patterns point to a public that is trying to make sense of threat before it can make sense of control measures.
When those questions go unanswered, misinformation fills the space. The analysis across six Ebola outbreaks shows that this is not an isolated problem tied to one event. It is a recurring feature of outbreak communication, and it becomes more dangerous when violence and mistrust are already in play.
Limits of the data and the unanswered operational question
The study’s strength is its multi-outbreak design and its use of two complementary digital data streams. Its limit is equally clear: search data are a proxy for public interest, not a direct measure of beliefs, behavior, or infection risk. Google users do not represent every affected community, and search language does not capture offline rumor networks in full.
Even so, the signal is strong enough to inform response planning. The remaining operational question is how quickly public health agencies can translate these digital signals into locally tailored messaging before conspiracy content hardens into resistance. In outbreaks where treatment centres are attacked and response teams are threatened, delay is not a neutral choice. It leaves the communication field to rumor.