Using non-traditional data to date Congo's Ebola outbreak
EpidemiologyComments
The report mentions 500 suspected cases, but it is unclear how the researchers distinguished Ebola symptoms from other endemic hemorrhagic fevers using only biographies and messages. Was there a specific keyword filter or a cross-reference with later clinical data?
official dates usually track the arrival of international aid, not the virus.
To expand on that, the discrepancy usually stems from the case definition used by the WHO versus local observations. Local records capture community-defined cases, which often precede the clinically confirmed cases required for an official declaration.
This is a common pattern in rural infrastructure; the official timeline usually aligns with when the budget is released for the response. The actual crisis always starts months earlier in the village records.
This is so fascinating... it reminds me of how researchers used Google search trends to track flu outbreaks before the clinics even saw the spike... the social patterns are often leading indicators!
If funeral biographies are this effective, why are we still pretending that top-down bureaucratic reporting is the gold standard for early detection? Is the system intentionally blind to the community level?
I disagree that this provides better accuracy. It increases sensitivity for early detection, but using non-clinical data like funeral bios introduces significant noise and false positives.