How Big Could the 2026 Bundibugyo Ebola Outbreak Get? CDC's Scenario Projections

Summary

• An outbreak of Bundibugyo virus disease (BVD), a kind of Ebola disease, centered in Ituri province of the Democratic Republic of the Congo (DRC), was under way in mid-2026.
• CDC used a transmission model to project its growth over 3 months, under different assumptions about how many had died by May 24, 2026, and how many patients would be identified and isolated.
• Assuming 50 deaths by May 24, isolating 70% of patients left only about a 1 in 20 chance of more than 10,000 cases in 3 months. With only 20% isolated, the chance of more than 20,000 cases was 65%.
• Large-scale, rapid action is needed to keep what is already the largest known BVD outbreak from becoming one of the largest Ebola epidemics in history.

The outbreak

On May 15, 2026, the health ministries of the DRC and Uganda declared BVD outbreaks. The cause, Orthoebolavirus bundibugyoense, has no approved vaccine or treatment. BVD is a severe hemorrhagic fever, spread by direct contact with the body fluids of someone infected or someone who has died of it. By June 2, there were 378 confirmed cases (363 in the DRC, 15 in Uganda) and 63 confirmed deaths (62 in the DRC, one in Uganda).

How the model works

The model, adapted from one used for earlier hemorrhagic fever outbreaks such as Marburg virus disease in Ethiopia in 2025, simulates outbreaks one infection at a time — a branching process:

• Each simulated outbreak begins with one person infected from an animal source (a spillover), who infects a random number of others according to the basic reproductive number (R0) — the average number of people one case infects in a susceptible population — and so on, generation by generation.
• A simulation ends when a generation infects no one, or at 5,000 deaths.
• Timing — infection to symptoms, symptoms to death or recovery — draws on earlier Ebola outbreaks, and BVD data where available. Simulated people are infectious from symptom onset, not after recovery, but can be after death.

Calibration: because the real death count was uncertain — DRC reports as of May 24 listed 10 confirmed and 223 suspected deaths — the model was fitted three times, to 50, 100 and 200 cumulative deaths by May 24, with the first death on or before April 24. For each, 500 matching simulations were kept.

Scenarios: from May 24, a set share of people with symptoms were isolated and treated, after an average delay of 2 days, and could infect no one; those who died in isolation were assumed safely buried (no washing or embalming, burial by trained teams in protective gear). Four levels were tested: 20% (poor), 50% (moderate), 70% (high) and 95% (extremely high), the last to estimate a floor. Each simulation ran to August 22 — 90 days of intervention.

Results

Assuming 50 deaths by May 24 — spillover most likely around February 19, 2026 (interquartile interval February 1–March 8):

Assuming 100 deaths — spillover around February 8 (January 21–February 27): at 20% isolation, 76% of simulations reached ≥20,000 cases and 87% ≥4,000 deaths; at 70%, 73% stayed under 2,000 deaths and 10% reached 4,000 or more.

Assuming 200 deaths — spillover around January 29 (January 9–February 18): the outbreak would already be larger when interventions began, so even at 70% isolation, 42% of simulations reached ≥10,000 cases.

Sensitivity: when R0 was above the median, outbreaks usually passed 10,000 cases and 2,000 deaths at 50% isolation or less, even assuming 50 deaths. At 70% isolation and 50 deaths, no simulation with R0 below the median reached 2,000 deaths, but 20% of those above it did. The fitted median R0 was 2.51 — nothing unusual for Ebola.

Simulated outbreak size by August 22, 2026, at each level of patient isolation. Figure from the CDC report.

What it means

Without large, sustained interventions quickly, this outbreak could grow as large as the 2014–2016 West Africa Ebola epidemic — more than 28,000 cases and 11,000 deaths — and under low isolation could become the second-largest Ebola outbreak in history within 3 months. Controlling it may take a response on the same scale as West Africa's.

• Very large outbreaks appeared even in the most optimistic calibrations when isolation was low.
• Fitting to more deaths was roughly like starting interventions later — so acting early lowers the odds of the worst outcomes.
• The big projections come mainly from how large the outbreak already was when first confirmed, not from an unusually high R0; for Ebola, the longer the gap between onset and detection, the bigger and longer the outbreak.

Risk to the U.S.: CDC still judges the risk to the general U.S. population as low. Even during the huge West Africa epidemic, only two people were infected in the United States — health care workers caring for a patient who had traveled before enhanced screening began at U.S. ports of entry — and both recovered.

Limits: the true death count to May 24 is unknown; R0 for Ebola varies widely, and good case and death data over time are needed to pin it down; the model leaves out behavior changes that would slow spread, immunity built up by infection (which could exaggerate projections over longer periods), and relapses after recovery.

What is needed

Urgent, sustained action: finding cases fast, tracing contacts, isolating and treating people with BVD, engaging communities, and safe and dignified burials — with resources that may rival the 2014–2016 response.

Sources

Based on Eric Q. Mooring, William T. Koval, Isobel Routledge and others, "Modeled Scenario Projections for the Ebola Disease Outbreak Caused by Bundibugyo Virus, 2026," MMWR, Centers for Disease Control and Prevention (Early Release, June 5, 2026); a work of the United States government in the public domain.