The Contact List Is Not the Outbreak

The often-repeated claim that 80% of cases had unknown chains confuses contact follow-up coverage with the share of cases linked to a chain.

Retrospective written 16 October 2026. Focus month: July 2026.

An outbreak map is made from names, dates, addresses, clinic visits and conversations. In July, an arresting claim circulated: 80% of Ebola cases came from unknown transmission chains. It sounded like a measure of surveillance failure. The WHO’s later review shows why careful denominators matter: about 80% referred to follow-up of listed contacts; only about 20% of cases were identified through that follow-up, and about 40% of cases were ultimately linked to a transmission chain.

Those are three different statements. One is the share of known contacts that response teams followed. Another is the share of cases found through following them. The third is how many cases investigators could connect to a chain. Turning the first into “80% of cases had unknown chains” inflates and mislabels the evidence. It also risks making a preventable information problem sound like a single catastrophic statistic.

What July showed

On 17 July, the WHO Regional Director reported 2,073 DRC cases and 796 deaths as of the previous day. The outbreak had expanded across health zones and provinces. The report described acceleration: confirmed cases were up 50% over two weeks, deaths 91%. These were source-reported counts, and the crisis was severe regardless of the precise contact-tracing denominator.

In August, the WHO Emergency Committee described the operational gaps more carefully: only around one in five cases had been detected through follow-up of listed contacts; around two in five cases were linked to chains; contact follow-up among those listed was around four in five. The difference between “contacts followed” and “cases found” is the central lesson. A contact list can be diligently worked and still fail to capture people whose exposure was never recorded, who moved, or who could not be safely reached.

The measurement problem

A case curve is not a transparent window onto infections. It records detected and reported cases. If laboratories expand, testing finds people who were previously invisible; if reports arrive late, cases may cluster on the publication date; if officials reconcile older records, a notification spike may not represent a sudden wave of new infections.

WHO’s later outbreak notices explicitly warn that a large reporting increase in July included data reconciliation, while also stating that most of the rise reflected expansion of the outbreak. Both things can be true: improved measurement can raise counts, and transmission can still be worsening. The honest response is to annotate the time series, not discard it or treat every rise as fresh infections.

Surveillance is also a relationship. The person who names a contact must trust that tracing will lead to support, not stigma. A patient should be able to seek care without losing a livelihood or being treated as a threat. A mobile reporting tool can help record symptoms or alert a clinic, but data collection without a response offer is likely to erode trust. Privacy safeguards are not a luxury during an outbreak; they help people participate.

The Pawanax lens

For Kenya and East Africa, the practical design goal is to shorten the path from an unusual signal to a useful action. That can include a clinic prompt for a relevant travel history, a referral to testing, a community health worker follow-up, or a county-to-national escalation. The system should record when a signal was raised, whether anyone acted on it, and what happened next. Counting messages sent is not the same as finding cases; counting contacts listed is not the same as reaching them.

The October 5 imported Bundibugyo case in Kenya makes the distinction immediate. WHO reported the case with travel history from DRC. That is an imported case; it does not demonstrate local transmission. Border screening can catch some illness, but it cannot replace accessible care, testing and follow-up after travel. A person may be pre-symptomatic, travel for urgent reasons, or enter through routes where screening is not possible.

The counterevidence is that surveillance can improve outcomes. WHO described cases identified in Europe being isolated after travelers reported symptoms and health authorities initiated tracing; no sustained transmission followed. Information systems, trained staff and prepared facilities can convert a detection into containment. The chain depends on all of them.

The next 90 days, seen from October 16

The 16 October forecast register carries F7: 20% for Kenya confirming at least one locally acquired Bundibugyo case by 31 January 2027. This is an inherited subjective judgment, uncalibrated and not validated by this retrospective. It will be scored against a confirmed classification from Kenyan or WHO authorities. The October imported case is not the event described by F7. The operational outlook is conditional: if surveillance detects illness early and access to testing and care is high, importations can remain isolated; if case finding misses community and clinic transmission, the chance of secondary spread rises. No unsupported numerical causal effect follows from the contact-tracing figures.

The July lesson is not that the system sees nothing. It is that a percentage travels badly when its denominator is dropped. In epidemic response, accurate language is part of accurate surveillance.

Part of The World Ledger: 2026, Read Backwards. AI-assisted research and drafting; linked sources and figures checked for this retrospective, published 16 October 2026.

Read the World Ledger

All 30 articles. Previous: The Budget Behind the Clinic. Next: The Settlement Is Part of the Map.

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: Contact tracing 2.jpg — CDC Global · CC BY 2.0. Photographed 14 August 2014 in West Africa during Ebola response. Historical illustrative example; not a current outbreak scene or Kenyan operation. Resized for web display.