An Outbreak Is a Map of Missing Names

As Bundibugyo Ebola expanded across provinces, the central operational question was whether surveillance could keep pace with the people it needed to find.

Retrospective, authored 16 October 2026. This article focuses on signals recorded in September; it does not claim these observations were published then.

The line on the map moves

A border is a line on a page. A journey to a hospital is a sequence of roads, buses, family homes and waiting rooms. Disease follows contact, not cartography. By September, the Bundibugyo Ebola outbreak in the Democratic Republic of the Congo had become a regional warning precisely because its geography was widening faster than any single health authority could observe it. WHO had already described intense transmission and geographic expansion in its 14 August update; the September report showed that the concern had persisted and the footprint continued to grow.

In its 25 September situation report, the World Health Organization recorded 7,890 confirmed cases and 3,799 deaths in the DRC as of 23 September, a crude case fatality ratio of 48.1 per cent. WHO said the outbreak had reached 63 health zones in seven provinces, and that 48 zones across six provinces had reported a case in the previous 21 days. The report also noted that 83.4 per cent of identified contacts had been monitored during the previous 24 hours. The WHO report contains both the scale of the response and the gaps it was trying to close.

That monitoring figure deserves care. It means roughly 16.6 per cent of the contacts due for follow-up were not recorded as seen in that 24-hour window. It does not mean that 16.6 per cent were infected, lost permanently or certain to transmit the virus. It points to workload, access, staffing and mobility problems that make a list difficult to complete.

WHO also described considerable variation among provinces. Ituri remained the epicentre; North Kivu's incidence had risen sharply in mid-September; newly affected zones appeared elsewhere. A national total can look like one curve even when local outbreaks are moving in different directions. That makes local surveillance essential: a fall in one province may coexist with a rise in another.

Contact tracing is care infrastructure

Tracing a contact is often described as a security operation. In practice, it is also care. A health worker must identify a person, locate them, explain why follow-up matters, build enough trust to ask about symptoms, and arrange care if illness begins. In settings affected by insecurity or displacement, that chain can break for reasons unrelated to public willingness: roads are unsafe, phone numbers change, people move for work, and clinics may be far away.

WHO's September report named several such barriers: conflict, insecurity, displacement and limited access to basic services. It warned that these conditions hamper case finding, tracing, infection prevention and timely care. The point is not that a phone app solves those conditions. It is that every tool should be designed around them. A missed entry might be a technical failure, a changed address, an unavailable phone or a family that does not trust the caller. Those explanations demand different responses.

There is a necessary counterweight to alarm. WHO's report also described a recent decline in the seven-day moving average after a resurgence earlier in September, and progress in monitoring contacts. The national trajectory was not a simple uninterrupted acceleration. Neither a high cumulative total nor one trend line tells us what will happen next week in each affected area.

The math: what a completion rate can and cannot say

At the reported moment, WHO counted 26,980 contacts seen out of 32,342 requiring follow-up. Divide 26,980 by 32,342 and the result is about 83.4 per cent. The remaining 5,362 are the difference between contacts requiring follow-up and contacts seen during that 24-hour period.

That subtraction is operationally useful: it gives response teams a queue to work. It is not an estimate of infections, because contact monitoring is not the same as diagnosis. It is also a snapshot, not a final tally of people never reached. A contact can be reached later, and a person not seen that day may or may not have symptoms.

This is why data quality needs a denominator and a timestamp. “We followed up 26,980 people” is meaningful only alongside how many required follow-up and over what interval. In a clinic, the parallel is a referral completion rate that includes all referrals made, not just those who returned. The numerator tells us activity; the denominator shows what remains.

A regional lens

The DRC shares borders and travel routes with countries whose residents move for trade, family and medical care. WHO noted that surveillance was active at official crossings but that travel also occurs along informal routes. By 16 October, Kenya's Ministry of Health had reported an imported Bundibugyo case. That later event should not be projected backwards into September; it does, however, illustrate why the September warning about cross-border spread was material. Kenya's Ministry announcement records the confirmed case and response measures.

For East African health systems, preparedness requires more than airport screening. It needs diagnostic capacity, trained staff, referral and isolation arrangements, consistent communication, and cooperation across borders. The WHO regional situation-report portal tracks updates and preparedness material. WHO also reported that there were no licensed vaccines or approved therapeutics specifically targeting Bundibugyo virus disease at that time, which increased the value of early detection and supportive care.

Pawanax builds tools used by clinicians and families, so a responsible digital layer should help route a concern to a human and maintain a record of what was asked and answered. It must protect privacy, avoid stigmatizing communities, and never turn an automated risk flag into a diagnosis. Surveillance only works when people will speak to it.

The next 90 days

F6 (80%) asks whether WHO's Bundibugyo PHEIC remains in force on 31 January 2027. F7 (20%) asks whether Kenya confirms a locally acquired, non-imported case by that date. Both are subjective, uncalibrated probabilities registered on 16 October. They will be scored against WHO's emergency status and Kenya's official case classification using the Brier rule (p − outcome)².

Those forecasts are deliberately distinct. A regional emergency can continue while Kenya prevents local transmission; an imported case does not by itself imply community spread. The next signal to watch is not only the case count. It is whether response teams can find contacts, provide timely care and explain the situation clearly enough that people keep coming forward.

Read the World Ledger

All 30 articles. Previous: The Price of Holding. Next: The Door Behind the Model.

Mathematical methods: 45 parameters, assumptions and limits. Download the research and executable model.

AI-assisted research and writing using parallel Luna agents; editorial synthesis under the APEX lens. Historical scenes are illustrative. Sources support the stated observations; forward scenarios and inherited probabilities remain conditional, uncalibrated judgments. Kenya Vision 2060 is a founder-authored thesis, not an adopted government plan.

Cover image: Ebola Virus Particles (7).jpg — NIAID · CC BY 2.0. Micrograph dated 17 June 2014; Commons does not give a specimen location. Illustrative microscopy, not a sample from the 2026 Bundibugyo outbreak. Resized for web display.