The doses that began arriving in the Democratic Republic of Congo on 22 August were not made for this outbreak. The international coordinating group on vaccine provision allocated 70,000 doses of Ervebo on 20 August, and 16,520 of them have landed. Ervebo is licensed against Zaire ebolavirus. The virus killing people in eastern Congo is Bundibugyo, a different species, for which no vaccine and no treatment has been approved anywhere. The World Health Organization has said plainly that it remains unknown whether Ervebo works against the Bundibugyo strain, and that early laboratory and animal data suggest it may provide some protection.
That is the honest position, and the allocation reflects it. Of the 70,000 doses, 20,000 are designated for a clinical trial and 50,000 for health workers. Read that split slowly. A fifth of the consignment is an experiment to find out whether the vaccine does anything at all against this virus, and the remainder is going to the people most likely to be exposed before the answer is known. The response to the largest Bundibugyo outbreak on record opens as a research protocol conducted at epidemic speed, because the alternative is to offer frontline staff nothing.
The staff figures explain why nobody waited. Julien Harneis, the United Nations senior Ebola coordinator, reported more than 160 health workers infected and 40 dead, against 5,021 cases and 2,325 deaths in figures published on 20 August. Health workers are not incidental casualties in an Ebola outbreak. They are the response capacity itself, and each infection removes a trained person from a system that has very few, while adding a treatment bed and a contact-tracing chain. Forty deaths among responders, in eastern provinces already short of clinicians, is a subtraction the response cannot make up while the emergency is running.
A fifth of the consignment is an experiment to find out whether the vaccine does anything at all against this virus, and the remainder is going to the people most likely to be exposed before the answer is known.
The structural point sits in the licence, not in the logistics. The world built a vaccine, and a stockpile behind it, in response to the West African epidemic of 2014 and the Congolese outbreaks that followed, all of them Zaire ebolavirus. That was a rational allocation of effort against the strain that had caused the emergencies. It also means the stockpile is a record of which outbreaks have already happened, and offers nothing certain against the one occurring now. Preparedness built strain by strain is preparedness that arrives one outbreak late by design.
Bundibugyo has been known since 2007. It has caused outbreaks in Uganda and in Congo before. It was not an unknown pathogen that emerged without warning, and the absence of a licensed countermeasure is not a scientific mystery. It reflects a development pipeline that moves when a strain kills at scale in a place that commands attention, and stops when the outbreak ends. The interval between outbreaks is when the work would have to be done, and it is exactly when nobody funds it.
The trial arm therefore carries weight beyond this epidemic. If Ervebo shows cross-protection against Bundibugyo, the global stockpile becomes usable against two species instead of one, and the next Bundibugyo emergency starts from a materially different position. If it does not, the finding is still worth having, because it ends the assumption that a stockpile exists for this virus and puts the gap where policymakers have to look at it. Either way the evidence will be generated in eastern Congo, by Congolese health workers, in a conflict-affected region, during one of the largest Ebola outbreaks on record. The knowledge produced will be used everywhere. The risk of producing it is carried in one place.
Nothing about the allocation is wrong. Sending 70,000 doses of an unproven but plausible vaccine to people being killed at this rate is the correct call, and the candour about what is and is not known is better than the alternative. The failure sits upstream, in the years when Bundibugyo was a known species with no commercial constituency and no emergency to force the question. Congo is now running the study that should have been completed before the outbreak began, and it is running it with the people it can least afford to lose.
