Search behavior during Ebola outbreaks exposed a sharp language gap in public information demand
The 2026 Bundibugyo virus disease outbreak in the Democratic Republic of the Congo and Uganda has unfolded alongside a second emergency: a widening information crisis. In that setting, investigators examined what people actually searched for across six Ebola outbreaks from 2014 through 2026, using Google Autocomplete and Google Trends to map public information demand in English and French. The approach matters because outbreak communication is often designed around what health officials assume people want to know, not what they are actively typing into search bars when fear rises and rumors spread faster than laboratory confirmation.
Across 83 seed terms, the autocomplete extraction generated 7,347 unique queries. The search landscape was dominated by questions about risk and transmission. People were not searching in the abstract. They were asking whether the virus could spread through touch, whether symptoms meant infection, and how likely exposure really was.
One signal stood out immediately: misinformation and search interest moved together.
Misinformation searches rose in step with overall Ebola attention
Temporal analysis across six outbreaks showed moderate to strong co-movement between misinformation-related search terms and overall Ebola search interest, with Spearman ρ ranging from 0.359 to 0.697 and nominal p values below 0.001. In practical terms, misinformation did not sit on the margins of public attention; it traveled with the outbreak itself, gaining visibility as case counts, media coverage, and community anxiety intensified. During an event like Bundibugyo virus disease, that linkage is clinically relevant because rumor does not remain online for long. It shapes care-seeking, trust in responders, and whether families accept isolation, treatment, or vaccination messaging.
The study also shows why epidemic communication cannot rely on generic templates. Search demand clustered around the basic mechanics of contagion, which suggests that many people were still trying to answer elementary but urgent questions: How is Ebola transmitted? What counts as exposure? Can symptoms be trusted? Those are not trivial queries. They are the first decision points in whether a person seeks care, self-isolates, or delays reporting contact with an infected individual.
French-language searches carried more fear and conspiracy content
The largest divergence emerged between the English-language and French-language query sets. French-language suggestions contained six times more misinformation and conspiracy content than English-language suggestions, 4.9% versus 0.8%. They also contained nearly nine times more fear-related content, 3.5% versus 0.4%.
Those differences are not cosmetic. They point to distinct information environments that may require different public health tactics, different messengers, and different vocabulary. In outbreak settings across central Africa, French is not merely a translation layer; it is often the language through which official instructions, media coverage, and rumors compete for legitimacy. When French-language searches carry a heavier burden of fear and conspiracy framing, the communication response has to account for tone as well as content. Reassurance that ignores local suspicion can backfire. Clinical facts delivered without cultural and linguistic fit can fail to land at the exact moment trust is needed most.
The asymmetry between language groups also suggests that the people most exposed to active transmission zones may not be receiving the same informational support as anglophone audiences. Public health systems often build risk communication around national or international English-language outputs, then assume translation is enough. This analysis argues otherwise. Translation does not erase the social meaning of words, and it does not neutralize the speed at which conspiracy narratives can spread through communities already living with conflict, displacement, or past grievances against health authorities.
Bundibugyo virus disease sharpened the stakes of message design
The current outbreak provided a particularly urgent backdrop because it is the first Ebola-related public health emergency of international concern caused by a species lacking a licensed vaccine. That absence raises the cost of poor communication. When there is no licensed vaccine to anchor a broad reassurance campaign, response teams have to depend even more heavily on rapid case finding, contact tracing, isolation, and community cooperation. Those tools only work when people believe the response system is credible.
The destructive effects of misinformation in Ebola outbreaks have long been visible in the field, where violence against response teams and destruction of treatment centres can follow rumor-rich environments. Search behavior offers a way to detect those pressures earlier and at scale. Google Autocomplete captures what the public is prompted to ask, while Google Trends shows how attention rises and falls over time. Together, they reveal not just curiosity, but the emotional and cognitive structure of outbreak awareness. A population asking predominantly about transmission is a population trying to decide whether danger is immediate. A population generating conspiracy-laden searches is one in which trust has already been strained.
This is where the study’s value becomes operational. Infodemic management is often treated as a communication add-on, but the data point in the opposite direction. During Ebola outbreaks, information itself becomes part of the transmission environment. If people search for fear, rumors, and conspiracy explanations in one language more than another, response teams need language-specific monitoring, language-specific counter-messaging, and community channels built around local trust networks rather than one-size-fits-all campaigns.
Why the search record matters for response teams on the ground
The most actionable insight is simple: public information demand is not uniform, and it is not politically neutral. English-language search activity and French-language search activity do not map onto the same emotional or misinformation profile, which means the same outbreak can generate two different communication crises at once. Health authorities, humanitarian groups, and digital surveillance teams can use this kind of analysis to identify which themes need immediate clarification, which myths are gaining traction, and where messaging should be adapted before rumor hardens into behavior.
In a disease cluster where distrust can cost lives, the online record is a real-time clinical signal of social risk. Search data cannot replace field epidemiology, but it can complement it by showing which questions are too urgent to ignore and which communities may need a different explanatory frame before the next contact tracing visit, burial intervention, or treatment referral.
The remaining problem is implementation. Search data can reveal linguistic gaps, yet the study does not answer how quickly health agencies can convert those signals into locally credible interventions during a fast-moving outbreak, or how well those interventions perform when violence, displacement, and historical mistrust are already in play.