The Emergency Nobody Priced

On 20 May 2026, the World Health Organization counted 53 confirmed Bundibugyo Ebola cases across the Democratic Republic of the Congo and Uganda. It also reported about 600 suspected cases and 139 suspected deaths. The gap between those numbers was not a forecast. It was a warning that the confirmed count described only the cases investigators had established, while the suspected total carried a different and less certain denominator. WHO’s briefing that day also said risk was high nationally and regionally, and low globally.

An emergency can be visible without being priced into the daily conversation. In May, oil, food and interest rates had market quotes. Outbreak readiness had no single number. Its costs sat in laboratories, clinics, transport, staff protection and public trust—systems whose value is easiest to see after they are needed.

The story that followed was not a smooth curve, and the data do not permit a clean line from one month’s count to the next. They do show how quickly the scale and geography of a public-health emergency can change, why “confirmed” and “suspected” must stay separate, and why a case count is not the same thing as a map of transmission.

What the first count meant

WHO reported 51 confirmed cases in the DRC and two in Uganda on 20 May, alongside the larger suspected totals. A suspected case is a person who meets a clinical or epidemiological definition and needs investigation; it is not interchangeable with a laboratory-confirmed infection. Counts can rise because transmission increases, because investigators test more people, because delayed samples are processed, or because reporting catches up. They can fall for the opposite reasons as well as because spread is slowing.

The distinction is more than technical. Treating confirmed cases as the whole outbreak can understate need; treating every suspected report as confirmed can mislead. A sound account gives both, labels them and states the date. It does not turn the suspected-to-confirmed gap into a claim about how many transmission chains are untracked. Those are different measures.

WHO said the risk was high in the affected countries and region, but low globally at the time. That assessment did not mean “no danger”; it described the geographic probability and potential consequences considered by the agency. Nor did it imply the outbreak was destined to become a global epidemic. Risk assessments are conditional judgments, revised as the evidence changes. The May briefing said the WHO Director-General had declared a Public Health Emergency of International Concern, a formal international status that calls for coordinated attention. It was not a prediction of where the next case would occur.

The response also faced a medical constraint: at the start of the outbreak, there was no licensed vaccine or specific therapeutic for Bundibugyo virus disease. That did not mean there was no care. Clinical support, infection prevention, diagnostics and measures to interrupt transmission remained central. The distinction between “no licensed product for this virus” and “nothing can be done” matters to patients and public confidence. WHO’s May briefing described the response needs; later vaccine research is not evidence that a licensed vaccine was available then.

The numbers grew, but they did not tell one simple story

By 1 July, WHO reported 1,460 confirmed cases and 452 deaths in the DRC across 36 health zones. Ituri accounted for 1,333 cases, 91.3 per cent of the national confirmed total at that date. The same 3 July situation report said teams had identified 10,821 contacts in Ituri and North Kivu and followed up 8,954. Those figures describe people on identified contact lists. They do not establish what percentage of all infections had been found, or what share of every transmission chain investigators could link.

WHO also warned that newly reported cases and deaths could include a backlog of samples and might not represent infections acquired on the date they appeared in the report. This is a basic feature of outbreak surveillance: infection, symptom onset, clinic visit, specimen collection, laboratory confirmation and public reporting happen at different times. A chart by report date mixes transmission with the speed and reach of the reporting system.

By the 23 October update, WHO said the DRC had recorded 9,704 confirmed cases and 4,687 deaths through 20 October, across 66 health zones in eight provinces. The outbreak had crossed a much wider area than it had in July. That later total is evidence of a larger known outbreak, not proof that every period within it had the same growth rate. Definitions, testing, affected locations and the backlog of reports can change. Calculating a single doubling time across months and presenting it as a forecast would give false precision.

The death-to-confirmed-case ratio is also easy to misuse. Dividing 4,687 by 9,704 gives about 48 per cent, but this crude ratio is not a patient’s probability of death. Cases and deaths may be reported at different times; people with recent diagnoses may not yet have known outcomes; unconfirmed infections are absent from the denominator. It is a descriptive ratio from a dated surveillance report, not a settled clinical fatality estimate for every patient or setting.

A border case is not the same as uncontrolled spread

Kenya’s first reported case in October made preparedness a live test beyond the DRC. The imported index patient travelled from the DRC through Uganda and reached Nairobi on 3 October, according to WHO’s regional office. In its 23 October update, WHO reported nine cases in Kenya, including one death, as of 20 October. It said that the eight cases after the imported index case were known contacts and that it had no evidence of wider community transmission at that cutoff. WHO’s update is careful about both the count and the limit of what it shows.

“No evidence of wider community transmission” describes information available by a date; it does not prove transmission could never occur. “Eight cases were known contacts” does not mean every exposure was identified. But an imported case alone does not establish widespread transmission in a new country. Contact tracing and testing help distinguish a linked cluster from broader spread.

WHO’s outbreak-control list includes rapid identification and investigation, isolation and care, contact tracing, safe burials and community engagement. Early symptoms can resemble other febrile illnesses, and laboratory testing is needed to confirm and distinguish Bundibugyo disease. The clinic is therefore part of the surveillance system: a patient needs to reach care, a provider needs a safe way to assess risk, a sample needs transport and testing, and results must reach the response team. Each handoff can be slowed by distance, insecurity, staff shortages or distrust.

That is why a “small” number can still pose a large operational challenge. A clinic needs space to separate a suspected patient, protective equipment, trained staff and a referral route. A laboratory needs capacity to test promptly; community teams need trusted channels. These are recurring costs, not supplies that arrive instantly after headlines.

There is a countercase to the view that the system had no buffers. Uganda reported no new case after 21 June in WHO’s early July account, and most of the listed contacts had completed their 21-day follow-up by 2 July. That shows why a defined cluster, intensive monitoring and adequate resources can make a difference. It does not guarantee the same result across the DRC’s larger geography, where insecurity and humanitarian conditions constrained access to some communities and facilities.

What “nobody priced” leaves out

Markets are not the only institutions that price risk. Governments and health agencies make budgets; hospitals decide whether to maintain reserve supplies and backup capacity; international organizations allocate response funds. But preventive spending competes against needs that are visible now. A functioning laboratory, an infection-control team or an available ambulance may look like idle capacity in a quiet month. During an outbreak, that slack can determine whether the first case is isolated quickly or found after several clinical visits.

The benefits are unevenly distributed. A well-staffed clinic with a reliable laboratory route may test a suspected case sooner. A remote community facing insecurity or long travel times may experience a longer delay and a greater burden on families. The ability to respond, however, differs sharply by place.

Emergency readiness cannot mean unlimited spare capacity for every rare threat. Staff, budgets and facilities are finite. The answer is not to call every risk imminent, but to fund capabilities used every day as well as during outbreaks: diagnostics, infection prevention, referral, trained staff, community communication and transparent surveillance.

That argument also depends on trust. People who fear stigma or coercive treatment may avoid clinics or withhold contact information. A response that explains what is known, what is still uncertain and what care is available can make reporting more credible. A sensational claim may attract attention for a day and damage cooperation for longer. Respectful communication is not decoration around the medical work; it is part of the work.

Forecasting without pretending to know

Forecasts are useful when they define an event and a date before the outcome is known. A register published on 16 October assigned an 80 per cent subjective probability to the WHO emergency still being in force on 31 January 2027, and 20 per cent to Kenya confirming a locally acquired Bundibugyo case by that date. These were uncalibrated judgments, not epidemiological model outputs. The register records the events for later scoring.

For a binary event, the Brier score is the squared difference between the stated probability and the outcome, coded as one if it occurs and zero if it does not. A probability of 0.80 scores 0.04 if the event occurs and 0.64 if it does not. That makes the forecast accountable, but one result cannot prove skill. Calibration is assessed across many forecasts, with carefully defined events and comparable time horizons.

The early outbreak did not make the later scale inevitable. It made the costs of uncertainty visible: confirmed and suspected counts diverged, contact lists were incomplete by definition, and the time between infection and public reporting mattered. By October, the DRC’s confirmed count had risen sharply, while the Kenyan update described a cluster linked to an imported case and no evidence of wider community transmission at that date. Both facts belong in the same account. The emergency was serious; its trajectory was not foregone.

No one can assign a single price to an early test, a prepared clinic or a trusted contact tracer. But leaving those capacities thin has costs: delayed diagnosis, greater exposure, harder logistics and families asked to navigate uncertainty without clear information. A sound forecast records uncertainty. A sound health system prepares without telling the public that the worst outcome is certain.

An Ebola isolation ward prepared in Lagos, Nigeria, photographed 19 August 2014. Historical file image from an earlier outbreak response; it does not depict a 2026 Bundibugyo treatment facility. Credit: CDC Global / Bryan Christensen, via Wikimedia Commons. Image source · CC BY 2.0 Generic