There is a number buried in WHO's most recent Ebola situation report that has received almost no attention in Western coverage of the outbreak. It is not the case fatality ratio, though that number is alarming enough. It is not the cumulative death toll, though crossing 2,000 deaths is a grim milestone by any measure. The number is 86.
Eighty-six days. That is how long it took the Bundibugyo virus disease outbreak in the Democratic Republic of the Congo to go from official declaration — May 15, 2026 — to 2,000 confirmed deaths. As of the data cutoff of August 9, 2026, WHO's situation report records 4,381 confirmed cases and 2,011 confirmed deaths, for a case fatality ratio of 45.9%.
I wrote about this outbreak in late July, when the reported CFR was 39%. The number has since moved up by nearly seven percentage points. That movement is the story — and understanding why requires pulling apart what a CFR actually measures, what it doesn't, and why the headline number is simultaneously too alarming and not alarming enough.
A CFR Is a Fraction. Both Halves Are Broken.
Case fatality ratio is deaths divided by confirmed cases. Simple arithmetic. The problem is that both the numerator and denominator are measured with error, and in an active outbreak in a conflict-affected region, those errors are not random — they are systematically biased in ways that make the ratio hard to interpret.
Start with the denominator: confirmed cases. Confirmation requires testing. Testing requires that sick people reach a health facility, that the facility has diagnostic capacity, and that the result gets reported into the surveillance system. In Ituri province, which WHO reports accounts for 85.8% of cumulative cases and 80.6% of cumulative deaths, active conflict and displacement have repeatedly disrupted health infrastructure. People who die at home, in transit, or in communities with no testing access are not confirmed cases. They may not even be suspected cases in the official count.
This means the denominator is almost certainly an undercount. And when you undercount the denominator while the numerator — deaths — is somewhat more likely to be captured (because deaths are harder to hide than mild illness, and community death reporting is often more robust than case detection), the CFR inflates. A true CFR of, say, 30% could easily appear as 45% if you're capturing 70% of deaths but only 50% of cases.
Now look at the numerator. Confirmed deaths are deaths in confirmed cases. But Bundibugyo virus disease, like other filovirus diseases, has a spectrum of severity. Mild and moderate cases are more likely to go undetected; severe cases that end in death are more likely to reach a facility and get tested. This severity bias further inflates the apparent CFR relative to the true infection fatality rate across the full population exposed.
None of this means 45.9% is fabricated. It means 45.9% is the CFR among confirmed cases — a specific, bounded measurement — and extrapolating it to "nearly half of everyone infected will die" requires assumptions about case ascertainment that the current surveillance system cannot support.
The 86-Day Number Is More Reliable Than the CFR — And More Frightening
Here is why the speed-to-2,000-deaths figure deserves more analytical weight than the CFR: it is less sensitive to the case detection problem.
Deaths are the more reliably captured end of the surveillance data. Community death reporting, burial surveillance, and retrospective case investigation all tend to capture deaths at higher rates than mild or moderate illness. So while the CFR is distorted by denominator undercounting, the raw death count — and the rate at which it is accumulating — is a more stable signal.
Eighty-six days to 2,000 confirmed deaths means roughly 23 confirmed deaths per day on average since the outbreak was declared. But the trajectory is not linear. WHO's situation report shows that in the single week ending August 9, there were 304 confirmed deaths — that is 43 per day, nearly double the outbreak average. The outbreak is not plateauing. It is accelerating.
The week-over-week data is the number epidemiologists watch when they want to know whether an outbreak is being controlled. A declining weekly case count suggests containment is working. A rising one suggests it is not. The jump from the outbreak's average daily death rate to the most recent week's rate is a signal that transmission chains are expanding faster than the response is closing them.
The Geographic Concentration Creates a False Sense of Containment
Ituri accounting for 85.8% of cumulative cases sounds, at first read, like good news. The outbreak is concentrated. Concentrated outbreaks are easier to ring-fence. This framing has appeared implicitly in coverage that treats the DRC outbreak as a regional crisis rather than a global health emergency.
The problem is that Ituri's concentration is partly an artifact of where surveillance capacity exists. The province has been the epicenter of conflict and displacement for years, which paradoxically means it has more international humanitarian presence — and therefore more testing infrastructure — than some surrounding areas. High case counts in Ituri may reflect better detection as much as higher transmission.
More importantly, WHO's report notes the outbreak also involves Uganda, which is why the situation report is jointly titled for both countries. Cross-border transmission is already documented. The 14.2% of cases outside Ituri represent geographic spread that is, by definition, harder to surveil and contain than the epicenter.
A CFR calculated on Ituri cases may differ from the CFR in areas with weaker health infrastructure. We don't know, because the data isn't broken out that way in the public situation report. That gap in the published data is itself informative.
What the Polling Methodology Debate Teaches Us About Outbreak Data
This is a useful moment to step back and apply a framework that comes up repeatedly in statistical journalism — the one Reuters uses when explaining its own polling methodology. When Reuters/Ipsos conducts a poll of at least 1,000 people, they report a margin of error of roughly 3 percentage points for the full sample, with explicitly higher margins for subgroups. They weight the data to account for known demographic biases in who responds.
Epidemiologists do something analogous when they try to estimate true infection rates from confirmed case data. They apply correction factors for case ascertainment — essentially, they try to weight the raw count to account for who is and isn't being detected. The DRC outbreak data, as published, does not include those correction factors in the headline CFR figure. The 45.9% is the raw ratio, unweighted for detection bias.
This is not a criticism of WHO's reporting — the situation reports are transparent about what they're measuring. It is a criticism of how that number travels once it leaves the situation report. By the time 45.9% appears in a headline, the methodological context has been stripped away, and readers are left with a number that sounds precise but is doing more interpretive work than it can support.
The Reuters polling explainer exists precisely because the outlet understands that a number without its methodology is a number that will be misread. Outbreak CFRs need the same treatment.
The Comparison That Actually Matters
The 2014–2016 West Africa Ebola outbreak — caused by Zaire ebolavirus, a different species than Bundibugyo — ultimately recorded a CFR of roughly 40% among confirmed cases, though estimates of the true infection fatality rate varied considerably based on serological surveys conducted afterward. The Bundibugyo species has historically shown lower CFRs in smaller outbreaks: the 2007 DRC outbreak recorded a CFR of approximately 25% across 56 confirmed cases, and the 2012 Uganda outbreak recorded roughly 25% across 24 confirmed cases.
Those historical figures come from outbreaks with case counts in the dozens, not thousands. Small-sample CFRs are notoriously unstable — a handful of additional deaths or recoveries can swing the ratio by ten percentage points. The current outbreak, with 4,381 confirmed cases, has a larger sample base, which makes the CFR estimate more statistically stable in one sense. But the detection bias problem scales with outbreak size: a larger outbreak in a more disrupted region means more missed cases, not fewer.
The current 45.9% CFR is higher than historical Bundibugyo benchmarks by a substantial margin. That gap could reflect genuine differences in outbreak severity, healthcare system capacity, or treatment access. It could also reflect worse case ascertainment — more missed mild cases — in a larger, more geographically dispersed outbreak. The data as published cannot distinguish between these explanations.
That is not a comfortable place to land. But it is the honest one.
What to Watch in the Next Two Situation Reports
WHO's situation reports are published weekly. The next two — covering data through August 16 and August 23 — will be the most informative yet, for a specific reason: the STAG-NTD meeting scheduled for August 24–25 at WHO headquarters will include a review of the mid-term evaluation of the WHO road map for neglected tropical diseases 2021–2030. Bundibugyo virus disease sits at the intersection of the NTD framework and the emergency response structure. Whether the STAG-NTD meeting addresses the outbreak directly or treats it as outside its remit will itself be a signal about how WHO is categorizing the response.
Three specific numbers to track in the coming reports: the weekly confirmed death count (to see whether the 304-in-one-week figure from August 9 was a peak or a new baseline), the percentage of cases outside Ituri (to track geographic spread), and any reported data on healthcare worker infections (a leading indicator of health system stress that often appears in situation reports before broader transmission acceleration becomes visible in the aggregate numbers).
The CFR will keep moving. It is the least stable number in the report and the one most likely to be quoted in headlines. Watch the weekly death velocity instead. That is the denominator that tells you whether this outbreak is being controlled.
