Estimating the current size of the 2026 DRC Bundibugyo virus outbreak

Authors: Sam Abbott and Sebastian Funk.

Last updated: on each rebuild. This is a live report, re-run as new data arrive, so the estimates change between updates.

Data as of: the most recent situation report. DRC counts come from the situation reports of the Institut National de Santé Publique (INSP). Uganda imports come from WHO. The rendered report fills in the build date and the exact data cut-off automatically.

Start here: summary dashboard · one-week-ahead forecast · limitations.

Full results: national · by province · forecasts · forecasts by province · in-sample checks · forecast evaluation.

Go deeper: methods · how this differs from McCabe et al. · reproduction number over time · model against observed · forecast skill against a baseline · frozen-fit evaluation · how the data streams compare · estimate evolution across releases · estimates against McCabe et al. · aim and origins.

Abstract. An outbreak of Ebola disease caused by Bundibugyo virus (BVD) is ongoing in the Democratic Republic of the Congo (DRC), with cases also detected across the border in Uganda. This is a real-time joint Bayesian estimate of the current size of that outbreak, refreshed as new data arrive. Most infections are not yet reported, so the current size has to be inferred from the surveillance data that are available. The model is a discrete-time renewal process on a daily grid, a meta-population over four provincial patches coupled by importation. It fits in a single posterior the daily counts the INSP situation reports publish, including the suspected and confirmed case and death series as well as the laboratory, isolation, treatment-centre and recovery records. We also digitise the epidemic curve by symptom-onset date as intermittently reported in the INSP situation reports and fit them using a model that accounts for time-varying ascertainment and reporting delays. Finally, we fit to data on initial exports of cases and deaths to Uganda, taken from the WHO situation reports and Disease Outbreak News. The same infection process generates all of them, staged to daily symptom onsets and routed into every observation stream. A second stage splits each patch across its health zones, fitted to the per-zone confirmed cases and deaths jointly as a composition within the patch. It conditions on the province fit's posterior weekly patch infections and does not feed back into it. A genetic bound on the time to the most recent common ancestor and priors from the McCabe et al. report complete the inputs. From these it estimates the infections and deaths to date, reported and unreported, the time-varying reproduction number with its growth rate and doubling time, the case-fatality ratio, and the ascertainment of each surveillance system. Every release projects each DRC stream a week ahead. The forecasts are scored with the continuous ranked probability score against the data that arrive later and against a persistence baseline. The aim is a transparent estimate of how far the outbreak has already grown, and of how much each published stream contributes to it.

Scope. This work adds an external view of the current situation, drawing on our understanding of real-time infectious disease dynamics and the infection process behind the observed surveillance counts. We are developing it and encourage feedback, so please get in touch. We support reuse and adaptation. Find out more in the contributing guide.

Use of AI. The model code and analysis were drafted by a language model, then reviewed and revised under human oversight. The named authors are responsible for that oversight. Every release is signed off by a person, who cuts it with an @release comment that needs write access to the repository.

Draft paper (work in progress)

The draft paper describes how this work developed, the current model, its evaluation and the lessons learnt. It is a work in progress, has not been peer reviewed and will change.

Installing the package

To use the model and the bundled outbreak data from your own Julia environment, add the package from the repository:

using Pkg
Pkg.add(url = "https://github.com/epiforecasts/BVDOutbreakSize")

You can then load the model machinery and the data the report is fitted to:

using BVDOutbreakSize
obs = load_observations()

This gives you the exported model components, constants and data loaders, enough to build your own analysis on top of the package. Reproducing the full report (fitting the models and writing the result tables and plots) can be done in a few ways, described next.

Running

There are a couple of ways to re-fit the model.

Re-fit from a clone

git clone --recurse-submodules https://github.com/epiforecasts/BVDOutbreakSize
cd BVDOutbreakSize
julia --project=. -e 'using Pkg; Pkg.instantiate()'
julia --project=. scripts/run.jl

scripts/run.jl fits the models and writes the output CSVs (the analysis literate is also run as part of the docs build). Running docs/pages/estimates/national.jl directly instead steps through the full narrative.

Re-fit without cloning

scripts/reproduce.jl fetches the package, instantiates its environment, and runs the fit:

curl -fsSL https://raw.githubusercontent.com/epiforecasts/BVDOutbreakSize/main/scripts/reproduce.jl | julia

Outputs land in ./bvd-output. Set BVD_OUTPUT_DIR to write them elsewhere, or BVD_REF to a release tag to reproduce a specific version. The script clones into a temporary directory and runs from there, so it leaves your own Julia environments untouched.

Render the docs page

Executes the literate and produces HTML at docs/build/:

julia --project=docs -e 'using Pkg; Pkg.develop(PackageSpec(path=pwd())); Pkg.instantiate()'
julia --project=docs docs/make.jl

Updating the data

The observations live in data/observations.toml. The literate picks up new numbers automatically. The figures come from the INSP situation reports, transcribed by INRB-UMIE/BDBV2026-Data, and are refreshed for each new sitrep with the scripts/ tooling (task download-sitreps, task confirm-data):

• The cumulative confirmed-case and confirmed-death series are read from the situation-report PDFs like every other stream, and cross-checked against the upstream national_* daily CSVs by scripts/confirm_insp_data.jl.
• Every other stream (suspected totals, laboratory cumulatives, the 24h analysed volume, daily new suspects, isolation occupancy, bed capacity and recoveries) is read directly from the situation-report PDFs (fetched by scripts/download_sitreps.jl into data/sitrep_pdfs/) and recorded in data/insp_sitrep_scanned.csv.

scripts/confirm_insp_data.jl cross-checks the scan against the upstream national series and exits non-zero on any disagreement it does not already document.

Outputs and releases

Each push to main regenerates the model outputs as part of the documentation build and publishes them as a GitHub Release. The latest release bundles: the saved result tables and plots; thinned posterior draws; a copy of the input observations.toml that produced them; and a site.zip snapshot of the rendered report site (download the latest). The README.txt inside says how to view it offline. The same artifacts are written to the repository's output/ directory on each build. Browse all releases for earlier output bundles. Major versions of the report are kept as GitHub Releases.

The rendered report is published from the gh-pages branch, where past and development versions of the analysis page can be found.

Submodules

• external/bdbv-linelist-analysis — Bayesian reanalysis of the 2012 Isiro BDBV line list (Rosello et al. 2015). Source of the informative onset-to-death gamma shape and scale priors.

Citation

If you use this project, please cite it and the situation reports that supply its data:

• This project — Abbott, S. and Funk, S. (2026). BVDOutbreakSize: estimating the current size of the 2026 DRC Bundibugyo virus outbreak. https://github.com/epiforecasts/BVDOutbreakSize. DOI: 10.5281/zenodo.20312758.
• INSP situation reports, the source of every DRC series — Institut National de Santé Publique, République Démocratique du Congo (2026). Situation reports on the 17th Ebola epidemic. https://insp.cd/ebola-17eme-epidemie/.
• WHO situation reports and Disease Outbreak News, the source of the Uganda exports and their deaths — World Health Organization (2026). Ebola disease caused by Bundibugyo virus, Democratic Republic of the Congo and Uganda.

Acknowledgements

Kath Sherratt and Samuel Brand were authors of versions 1.0.0 to 3.0.0. We thank them for their contributions to those versions and their review of them.

Funding

This work was funded by the National Institute for Health and Care Research (NIHR) Health Protection Research Unit in Health Analytics & Modelling, a partnership between the UK Health Security Agency, Imperial College London and the London School of Hygiene & Tropical Medicine (grant code NIHR207404). The views expressed are those of the author(s) and not necessarily those of the NIHR, UK Health Security Agency or the Department of Health and Social Care.

Further references

The works this project builds on and compares against.

• McCabe et al. estimates that this work re-implements and compares against, in all three released versions (the two Imperial reports and the peer-reviewed Lancet publication) — McCabe, R., Ebbarnezh, L., Okware, S., Fotsing, R., Koua, E., Mbaka, P., Lofungola, A., van Elsland, S. L., McMenamin, M., Ferguson, N., le Polain de Waroux, O., Cori, A. (2026). Estimation of the size of the outbreak of Ebola disease caused by Bundibugyo virus in DRC. Imperial College London, 18 May 2026. DOI: 10.25560/130007. Report page. McCabe, R. and others (2026). Estimation of the size of the Ebola outbreak caused by Bundibugyo virus in DRC: May 20, 2026 update. Imperial College London, 20 May 2026. DOI: 10.25560/13005307. Report PDF. McCabe, R., Ebbarnezh, L., Okware, S., Fotsing, R., Koua, E., Mbaka, P., Lofungola, A., Ebengo, D. M., Mbala, P. K., Bishola, T. T., Ibolobolo, C. M., Matondo, H. M., Sibo, J.-C. M., van Elsland, S. L., McMenamin, M., Ferguson, N., le Polain de Waroux, O., Cori, A. (2026). Estimation of the Ebola outbreak size in the Democratic Republic of the Congo. The Lancet Infectious Diseases, correspondence, online first 9 June 2026; estimates as of 27 May 2026. DOI: 10.1016/S1473-3099(26)00299-9.
• Chamla et al. projection that this work compares against — a stochastic SEIRD ensemble that projects cumulative laboratory-confirmed cases forward (a confirmed-case count, not an ascertainment-corrected total) — Chamla, D., Belizaire, M. R. D., Co, I. F., Jinadu, A., Mamadu, I., Atagbaza, A. O. (2026). Size of the 2026 Ebola outbreak and risk of cross-border spillover from Bundibugyo virus in Ituri Province, DR Congo, and its implications for preparedness: a recalibrated stochastic modelling study. The Lancet Infectious Diseases, 25 June 2026. DOI: 10.1016/S1473-3099(26)00320-8.
• Onset-to-death delay reanalysis that this work uses for delay priors — Funk, S. (2026). bdbv-linelist-analysis: Bayesian reanalysis of the 2012 Isiro Bundibugyo line list. .