Scenario wins: lewinke-thinking-bot* (97) cassi (26) hayek-bot (22) AtlasForecasting-bot (21) Mantic (18) Panshul42 (7)
| Figure/Metric | Value | Source | Significance |
|---|---|---|---|
| Suspected Cases (Africa CDC) | 336 | Africa CDC (May 16) | Current high-water mark for resolution criteria. |
| Suspected Cases (WHO) | 246 | WHO (May 16) | Conservative baseline for international reporting. |
| Suspected Deaths | 80-88 | WHO/Africa CDC | Indicates a high early mortality rate (~25-30%). |
| Lab-Confirmed Cases | 8-13 | DRC Health Ministry | Low number of confirmed vs. suspected suggests diagnostic lag. |
| Health Worker Deaths | 4 | WHO | High risk of nosocomial spread in informal clinics. |
| Policy Rate (BCC) | 13.5% | Central Bank of Congo | Reflects broader economic context/inflationary environment. |
My analysis of the 2026 Ituri Ebola outbreak focuses on the immediate reporting trajectory between May 17 and June 1, 2026. The baseline data shows a significant discrepancy: the WHO reported 246 suspected cases as of May 16, while Africa CDC reported 336 suspected cases for the same period. The resolution criteria specify using the most recently published figure from any reputable source. Consequently, I have treated the Africa CDC figure of 336 as the effective current floor for the prediction.
The outbreak is centered in Mongwalu, Rwampara, and Bunia health zones. Mongwalu is a peri-urban gold-mining hub with intense population mobility, which historically accelerates disease transmission. Furthermore, the detection of cases in Kampala and Kinshasa suggests that the virus has already exited the primary hotspot, increasing the likelihood of broader ‘suspected’ case reporting in new jurisdictions. The Bundibugyo strain (BDBV) is a critical factor; because there are no approved vaccines or therapeutics for BDBV, containment relies purely on surveillance, isolation, and contact tracing. This typically results in a longer tail of suspected cases compared to Zaire-strain outbreaks where the Ervebo vaccine can quickly dampen transmission.
To reach the upper bound of the provided range (445.5) by June 1, the outbreak needs to report approximately 7.3 new suspected cases per day from the 336 baseline. Given that the WHO noted “increasing trends” and many deaths remain under investigation, a growth rate of 10 to 15 cases per day is well within historical norms for the first month of an Ebola outbreak in the DRC. However, I have weighted the forecast toward the lower end of the 330–430 range to account for the possibility of reporting lags caused by insecurity in Ituri (linked to CODECO and ADF armed groups) and the potential for ‘suspected’ cases to be reclassified as ‘discarded’ rather than being added to a cumulative total. The mixture of scenarios considers a 40% probability of stagnant reporting (near 336), a 45% probability of moderate growth (reaching ~480), and a 15% probability of a surge (exceeding 600). The resulting 50% HDI of 323.50 to 422.50 reflects this balance between the aggressive initial reporting and the logistical challenges of maintaining high-frequency surveillance in a conflict zone.
Question: will the cumulative number of suspected Ebola cases in the Ituri outbreak exceed the upper bound (445.5) by the June 1 resolution date? As of mid‑May the Africa CDC reported 336 suspected cases and WHO declared a PHEIC on May 17; both forecasters expect rapid growth in suspected cases driven by aggressive contact tracing and enhanced surveillance. The outbreak is caused by the rarer Bundibugyo virus and local GeneXpert machines in Bunia aren’t calibrated for it, so samples must be shipped to Kinshasa—creating a laboratory confirmation bottleneck that will likely inflate the backlog of suspected cases. One forecaster notes a single‑day jump of 90 suspected cases, and both place the majority of probability mass above the 445.5 threshold (67% and 65%, respectively), with one concentrating much mass at the upper bound. They acknowledge a small but nonzero chance that scaled‑up testing or reclassification could reduce suspected counts, but overall both teams strongly agree the suspected count will very likely exceed the provided upper bound by June 1.
Forecast: Probability distribution:
Key drivers are (1) what number is used in the last reputable publication before June 1 (WHO vs Africa CDC/media figures have already conflicted), and (2) how much additional case-finding and transmission occurs over the ~15 days remaining. The WHO DON in the prompt gives 246 suspected as of May 16, but multiple forecasters report credible May 16–17 updates citing ~336 suspected, suggesting either rapid growth and/or definitional/reporting differences. Because the resolution uses the most recently published suspected-case figure (not necessarily monotone and not necessarily the maximum), there remains a small but real chance the latest published number before the deadline is still 246 (via the explicit fallback, or via a later publication repeating the older WHO figure). Most probability mass is placed on further increases into the high 300s/low 400s driven by intensified surveillance after the PHEIC plus ongoing spread in a high-mobility, insecure area. I still assign a meaningful upper tail beyond the open bound (>445) for scenarios with continued multi-zone spread and/or a backlog/reporting “dump” that pushes reported suspected cases above 445 before June 1.
An amateur forecast is likely to (a) anchor on 246 and ignore the possibility of later published higher counts, or (b) assume a single early jump implies near-certain explosive growth far beyond the short time window. This forecast explicitly accounts for the resolution mechanic (latest-published figure, possible definitional inconsistency, and the 246 fallback) while still incorporating realistic outbreak/reporting acceleration and preserving an honest open-upper tail.
Forecast rationale (numeric):
— Iteration 1 — Across the forecasts, the reasoning is anchored on the current reported baseline of 246 suspected cases as of May 15, then extrapolated forward to June 1 using a short time horizon of roughly two weeks.
The collective reasoning suggests a moderate upward increase from the current 246 suspected cases, with the most plausible endpoint by June 1 in the mid-to-high 300s, while recognizing substantial uncertainty from surveillance intensity, case definition effects, and the possibility of either stalled reporting or a sudden surge.
— Iteration 2 — Across the forecasts, the core reasoning is consistent: the reported suspected-case count is expected to rise from the most recent mid-May baseline because the outbreak is still active and post-PHEIC surveillance should intensify case finding over the following two weeks.
The forecasts collectively argue for a continued increase in suspected Ebola cases by June 1, driven less by new transmission assumptions than by improved detection and reporting in a setting with likely undercounting. The most likely outcome is in the upper 300s to low 400s, but the uncertainty is wide enough that substantially higher totals remain plausible.
— Iteration 3 — Across the forecasts, the main reasoning is anchored on two reported starting points: the WHO-confirmed baseline of 246 suspected cases on May 15 and the later Africa CDC figure of 336 on May 17. Most of the forecasting logic treats the May 17 count as the best near-term anchor, while also allowing for the possibility that some of the increase reflects backlog clearing or surveillance expansion rather than pure transmission.
The central expectation is that suspected case counts will continue rising through June 1, but only over a short 2-week window, so growth is expected to be meaningful but not explosive. The models generally assume a moderate daily increase, roughly in the range of 3–15 additional suspected cases per day, producing a consensus estimate in the low 400s, roughly around 400–420 total cases by June 1.
There is broad agreement that the outcome is likely to stay below the soft upper guidance of about 445.5, though several forecasts keep some probability mass above that level in case of rapid spread, improved detection, or retrospective case finding in insecure and hard-to-reach areas. The lower tail reflects possible downward revisions, reclassification, or slower reporting, which could keep totals closer to the earlier baseline than the latest count suggests.
Overall, the shared pattern is: start from the latest reported 336 cases, expect moderate near-term growth, center the forecast around the low 400s, and retain a wide uncertainty range because suspected-case counts can shift quickly with surveillance quality, access, and backlog effects.
The Core Dynamic The rationales universally agree that the final reported metric will be determined by a race between active epidemiological case-finding (which adds to the suspected count) and laboratory testing capacity (which clears and removes cases once tested).
Factors Driving Case Counts Upward Following the declaration of a Public Health Emergency of International Concern (PHEIC), an influx of funding and contact tracers will trigger intense active surveillance. To avoid missing early infections, health workers will apply a highly sensitive, broad clinical case definition. Because the region is hyper-endemic for other febrile illnesses like malaria and typhoid, this wide dragnet will inevitably capture a massive volume of symptomatic individuals, rapidly inflating the “suspected” list. Additionally, retrospective case finding could uncover historical cases, leading to sudden bulk additions to the data.
Testing Bottlenecks and Conflict Historically, suspected case counts drop as rapid molecular testing clears the backlog. However, Ituri Province is an active conflict zone severely impacted by armed militia groups. This insecurity, combined with poor infrastructure, creates severe logistical bottlenecks. Transporting physical blood samples from remote, transient mining areas to central reference laboratories will be dangerous and slow. Consequently, the rate of identifying new suspected cases is expected to vastly outpace diagnostic clearance, causing the pending testing backlog to swell.
Reporting Ambiguities and Scenarios Forecasters note that the final trajectory heavily depends on public health reporting practices—specifically, whether authorities report an actively fluctuating backlog of pending tests or a strictly increasing cumulative total of all investigated individuals.
Question: How many total suspected cases of Ebola will be reported in the 2026 Ituri Province outbreak by June 1, 2026? Final Prediction: Probability distribution:
Total Cost: extra_metadata_in_explanation is disabled Time Spent: extra_metadata_in_explanation is disabled LLMs: extra_metadata_in_explanation is disabled Bot Name: extra_metadata_in_explanation is disabled
Forecaster 1: Probability distribution:
Forecaster 2: Probability distribution:
Forecaster 3: Probability distribution:
As of May 16, 2026 the research reports 246 suspected Ebola cases (80 suspected deaths) and 8 laboratory-confirmed cases (4 confirmed deaths) in Ituri Province, DRC, with the virus identified as Bundibugyo ebolavirus. Cases are concentrated in Mongwalu and Rwampara health zones (with suspected cases in Bunia), and there are apparently unrelated confirmed cases reported in Kampala, Uganda and one in Kinshasa. The outbreak was officially confirmed May 15, 2026 (WHO learned of suspected cases on May 5); DRC’s Institut National de Recherche Biomédicale reportedly confirmed Bundibugyo virus in 13 of 20 samples. WHO declared a Public Health Emergency of International Concern on May 17, 2026.
For forecasting context the research compares this outbreak to historical reference classes and epidemiological parameters: prediction markets (Polymarket) gave low probabilities (about 11%) for a 2026 Ebola pandemic or a U.S. case by late June, historical Bundibugyo experience (Uganda 2007) had ~192 suspected cases over ~2–3 months, while larger DRC Zaire-strain outbreaks (2018–2020) reached thousands of cases but benefited from vaccines. Key epidemiological values cited include an average incubation ~12.7 days, serial interval ~12 days, and R0 typically 1.3–2.7 in Central African outbreaks. The research lists risk factors that could drive spread (no approved vaccines for Bundibugyo, urban and mining-area transmission, population movement, insecurity, poor infrastructure, high starting case count) alongside containment factors (rapid international response, DRC outbreak experience, surveillance activation). It notes surveillance dynamics that often produce early spikes in suspected-case counts and highlights the short accumulation window considered (about 14.5 days from the May 17 question opening to the June 1 resolution), with a baseline of 246 suspected cases on May 16, 2026.
Sources used in the research (as named in the rundown; original numbered references [1]–[33] were cited in the research but direct URLs were not provided): WHO (situation reports and PHEIC declaration), DRC Institut National de Recherche Biomédicale (laboratory confirmation), Polymarket (prediction market pages for “Ebola pandemic in 2026?” and “Ebola case in the US by June 30?”), Africa CDC, historical outbreak reports (2007 Bundibugyo Uganda), 2018–2020 DRC Kivu/Ituri outbreak reports, 2025 Kasai outbreak reports, 2022 Uganda Sudan virus outbreak reports, and literature on the 2014–2016 West Africa outbreak and Ebola epidemiology. No verbatim web links were included in the provided research document.
As of May 16, 2026, the outbreak statistics are:
The outbreak was officially confirmed on May 15, 2026, though WHO learned of suspected cases on May 5, 2026 [23]. Laboratory testing by DRC’s Institut National de Recherche Biomédical confirmed Bundibugyo virus in 13 of 20 samples [3][24].
Polymarket markets (as of May 15-17, 2026):
“Ebola pandemic in 2026?”: 11% probability, $39,268 trading volume [20]
Resolves Yes only if WHO explicitly declares an Ebola “pandemic” by December 31, 2026
Market consensus suggests global pandemic is highly unlikely despite current outbreak
“Ebola case in the US by June 30?”: 11% probability, $22,134 trading volume [21]
Low probability reflects traders’ assessment of minimal international spread risk
Based on limited cross-border travel, established surveillance, and containment protocols
These markets indicate forecasters view international spread as unlikely, suggesting expectations for containment within the region.
The only other known Bundibugyo outbreak provides the most relevant reference class:
2018-2020 DRC outbreak (Kivu/Ituri provinces, Zaire strain):
September-December 2025 DRC outbreak (Kasai province):
Other DRC outbreaks (2017-2025):
Transmission dynamics from historical outbreaks:
Effective population at risk: In the 2018-2020 DRC outbreak, modeling estimated ~16,000 individuals at risk despite 16 million demographic population, reflecting impact of ring vaccination, contact tracing, and quarantine [14][33]
Critical considerations for suspected case counts:
Historical patterns show that early in outbreaks, suspected case counts can increase rapidly as surveillance systems activate and contact tracing identifies additional cases, even if active transmission is being controlled.
The 2026 Ituri Province outbreak presents a complex forecasting challenge with the current baseline at 246 suspected cases as of May 16, 2026. The most relevant historical comparison—the 2007 Bundibugyo outbreak in Uganda—totaled approximately 192 suspected cases over 2-3 months, though that outbreak occurred in a different context. The current outbreak benefits from rapid international response but faces challenges from lack of approved vaccines, security issues, and high population mobility. Prediction markets assign low probability (11%) to major international spread, suggesting market participants expect regional containment. The short timeframe to resolution (14.5 days) and typical reporting patterns will be critical factors in determining the final count.
(a) Time left: About 14 days 11 hours from the question-open time of May 17, 2026 13:00 UTC to June 1, 2026 00:00 UTC.
(b) Outside view: The closest reference class is prior Ebola outbreaks in Central/East Africa, especially non-Zaire strains with no licensed vaccine. The 2007 Bundibugyo Uganda outbreak rose from 51 suspected cases on Nov. 28 to 93 suspected cases by Dec. 7, and ultimately infected roughly 149 people; the 2012 Bundibugyo DRC outbreak had 57 cases and 29 deaths; the 2022 Uganda Sudan-virus outbreak had 164 total cases. Larger conflict-zone DRC outbreaks can be far bigger: the 2018–2020 North Kivu/Ituri outbreak reached 3,481 cases, but over nearly two years and with a different strain/setting. (who.int)
(c) Inside view: This outbreak starts unusually high for a Bundibugyo event: WHO’s DON reports 246 suspected cases and 80 deaths as of May 15, while AP/Africa CDC later reported 336 suspected cases and 87 deaths by Saturday, May 16. That already pushes me above the question creator’s suggested “likely not higher than 445” anchor. Upward factors are: active community transmission, insecurity limiting surveillance, high mobility around Mongwalu mining activity, spread across Mongwalu/Rwampara/Bunia, and no licensed Bundibugyo vaccine or specific therapeutics. Downward factors are: rapid PHEIC declaration, DRC/WHO experience, immediate international coordination, and the fact that suspected-case growth may partly reflect backlog discovery rather than new transmission. (who.int)
(d) Outcome if nothing changed: If no further accepted update appears, I would expect the resolver to use 336 suspected cases, because AP reported Africa CDC’s updated suspected-case count after the initial 246 figure. If the resolver ignored that AP/Africa CDC update and relied only on the prompt/WHO baseline, the fallback would be 246. (apnews.com)
(e) Outcome if the current trend continued: A naive continuation of the reported increase from 246 to 336 in roughly one day would imply around 1,600–1,700 suspected cases by June 1. I treat that as an upper-trend scenario, not a central forecast, because the first jump likely includes intensified case finding and backlog classification.
(f) Expectations of experts and markets: WHO escalated to a PHEIC but explicitly not a pandemic emergency; Africa CDC/WHO statements emphasize serious spread risk but also a coordinated containment response. The research brief’s Polymarket references put broad Ebola-pandemic / U.S.-case probabilities around 11%, consistent with market expectations of regional containment rather than global spread. (who.int)
(g) Unexpected low scenario: Many suspected cases are later attributed to other febrile/hemorrhagic illnesses, new suspect reporting slows after the initial surveillance sweep, and the most recent pre-June-1 public figure remains near 336–400.
(h) Unexpected high scenario: Retrospective investigation uncovers large unreported chains in mining communities, Bunia, North Kivu, or travelers linked to Kinshasa/Uganda; insecurity blocks contact tracing; and broad suspected-case definitions pull in many alerts, pushing the reported suspected count above 2,000 before June 1.
Rationale:
Anchoring on prior Bundibugyo/Sudan outbreaks would normally put a full-outbreak total in the low hundreds, and a two-week-forward suspected count perhaps below 500. But the current event already has an unusually large suspected denominator and many deaths at confirmation, so I move sharply upward from that base rate. I do not simply extrapolate the +90/day jump because early outbreak reports are dominated by surveillance activation, retrospective case finding, and reclassification. My median is therefore a moderate-growth outcome: substantial further reporting, but not uncontrolled exponential growth. I put meaningful mass above 1,000 because “suspected cases” can grow quickly once contact tracing and alert systems expand, and because this outbreak has difficult terrain, insecurity, mobility, urban exposure, and no strain-matched vaccine.
Percentile 0.1: 246.1
Percentile 1: 300.1
Percentile 5: 336.49
Percentile 10: 390
Percentile 20: 445
Percentile 30: 520
Percentile 40: 600
Percentile 50: 700
Percentile 60: 830
Percentile 70: 1000
Percentile 80: 1250
Percentile 90: 1800
Percentile 95: 2500
Percentile 99: 5000
Percentile 99.9: 10000
I agree with the critic’s main point: the resolution mechanic is unusually important here. This is not a forecast of true cumulative infections; it is a forecast of the last reported suspected-case figure before June 1, 2026 00:00 UTC. That makes a stale-number / reporting-pivot scenario real. AP has a clean later figure of 336 suspected and 13 confirmed cases from Africa CDC, while WHO’s earlier Disease Outbreak News gave 246 suspected cases and 80 deaths as of May 15, and noted ongoing investigation of death clusters beyond the three initially affected health zones. (apnews.com)
I still defend the original upward skew. The current count is already high for a Bundibugyo outbreak, and AP reports Africa CDC describing “active community transmission,” high active community burden in Mongwalu, spread across Mongwalu/Rwampara/Bunia, insecurity, and weak containment conditions. WHO/AP also emphasize uncertainty about true magnitude and geographic spread, with suspected cases being investigated in additional areas. (apnews.com) So I do not want to move the entire distribution down to the creator’s “likely not higher than 445” intuition.
My adjustment:
I’m now centering a bit lower than my prior median, but keeping a larger high-end tail. My median is about 600 suspected cases, with substantial probability mass in the 336–450 range from reporting mechanics and a meaningful tail above 1,500 from surveillance avalanche / hidden-cluster scenarios.
Updated prediction:
Percentile 0.1: 246.05
Percentile 1: 300.49
Percentile 5: 336.05
Percentile 10: 336.35
Percentile 20: 390.49
Percentile 30: 465.49
Percentile 40: 535.49
Percentile 50: 605.49
Percentile 60: 710.49
Percentile 70: 850.49
Percentile 80: 1060.49
Percentile 90: 1420.49
Percentile 95: 2020.49
Percentile 99: 3850.49
Percentile 99.9: 7400.49
Conditioning on no annulment, and assuming no already-known event immediately resolves the question.
I think the critic’s strongest point is that my last update overcorrected downward in the center. The June 1 number will not mainly reflect new infections generated after May 17; it will also reflect retrospective case-finding from the late-April to mid-May detection gap, plus PHEIC-driven surveillance and media/agency updates. WHO says the first known suspected case had symptom onset on April 24, the outbreak was not laboratory-confirmed until May 15, and unusual compatible death clusters were still being investigated in Ituri and North Kivu. That argues for more upward movement than my previous median of ~640 allowed. (who.int)
I also accept the critic’s point that the low tail should be lumpy, not smooth. The official WHO/DON figure is still 246 suspected cases and 80 suspected deaths in Ituri, while AP reported Africa CDC’s later 336 suspected and 13 confirmed cases by Saturday, May 16. If no clean later suspected-case update appears, or if the resolver privileges WHO’s official figure, the outcome could stick near 246 or 336. (who.int)
Where I still do not fully follow the critic is the far-right tail. A suspected-case count in the several-thousand range by June 1 is plausible because surveillance could widen dramatically, but a 99.9th percentile above 20,000 feels too high for a two-week horizon under a resolution framed around the Ituri Province outbreak rather than all regional alerts. Ebola suspected-case reporting is broader than confirmed/probable reporting, but it is not simply every fever screened at a border. So I am fattening the right tail materially, but not as much as the critic’s 22,000 at 99.9%.
Net: I now put the median around 780, with meaningful point-mass-like probability at 246/336, and a much higher 90th–99th range than in my prior forecast. The most likely path is an Africa CDC/AP/WHO update in late May with a cumulative suspected count somewhere in the high hundreds to low thousands, driven by retrospective case-finding, community-death investigations, and broadened surveillance.
Updated prediction:
Percentile 0.1: 246.01
Percentile 1: 246.10
Percentile 5: 246.49
Percentile 10: 336.49
Percentile 20: 430.49
Percentile 30: 560.49
Percentile 40: 670.49
Percentile 50: 780.49
Percentile 60: 950.49
Percentile 70: 1220.49
Percentile 80: 1600.49
Percentile 90: 2350.49
Percentile 95: 3500.49
Percentile 99: 7800.49
Percentile 99.9: 17000.49
Agent Forecast (discrete, confidence: medium)
Base Rate: 2007-08 Bundibugyo Uganda outbreak: 192 total suspected cases over 6 months. 2007 DRC Kasai Zaire outbreak: 264 cases. Both are far smaller than the current outbreak which already has 336 suspected cases in week 3 of detection. The CDC defined Scenario B (100-500 cases) and Scenario C (500+ with exponential growth) for comparison; this outbreak is already in or approaching Scenario C territory.
Initial Reasoning:
Key Uncertainties:
Evidence Sources (8):
Grounding Facts (passed to reviewers):
Draft: 446+: 87.0%, 336: 0.2%, 439: 0.1%, 406: 0.1%, 404: 0.1%
Reviewer Panel: N/A
forecast_sdk | 339.2s | research $0.6763 = $0.6763
Extracted Percentile Values: {0.1: 248, 1.0: 280, 5.0: 320, 10.0: 350, 15.0: 375, 20.0: 395, 25.0: 410, 30.0: 425, 35.0: 440, 40.0: 455, 45.0: 470, 46.0: 473, 47.0: 476, 48.0: 479, 49.0: 482, 50.0: 485, 51.0: 490, 52.0: 495, 53.0: 500, 54.0: 505, 55.0: 512, 60.0: 540, 65.0: 570, 70.0: 600, 75.0: 640, 80.0: 680, 85.0: 730, 90.0: 790, 95.0: 900, 99.0: 1150, 99.9: 1600}
Model’s Answer:
This is a Cumulative Count Event forecasting question for the 2026 Ituri Province Ebola outbreak, measuring suspected cases through June 1, 2026.
Current Data Points (as of May 16-17, 2026):
The discrepancy between 246 and 336 likely reflects different data cut-offs and inclusion criteria. Given that multiple media outlets and the PHEIC declaration reference “more than 300 suspected cases,” the higher figure appears to be the more current operational count.
Factors pushing suspected cases higher:
Factors limiting growth:
Assumption 1: Updated figures will be published before June 1, 2026.
Assumption 2: The current true suspected count is approximately 330-340.
Assumption 3: International response will begin slowing transmission within the forecast window.
With approximately 14 days until resolution:
Historical benchmarks suggest:
Low outcome (10th percentile ~350 cases): If the current 336 figure represents near-peak surveillance detection, and rapid response effectively contains further spread, with minimal retrospective cases added. This would require unusually effective early intervention.
Central estimate (50th percentile ~485 cases): Moderate continued growth with surveillance expansion adding ~100-150 more suspected cases over 14 days as contact tracing and case-finding intensify, while transmission begins slowing due to PHEIC response.
High outcome (90th percentile ~790 cases): Continued rapid transmission in high-mobility mining areas, combined with aggressive surveillance detecting many additional cases and retrospective identification of earlier infections. Security disruptions could prevent effective response.
Very high outcome (99th percentile ~1150 cases): Multiple transmission chains establish in urban areas (Bunia, Kinshasa), overwhelmed response capacity, and significant under-detection being corrected through expanded surveillance.
Percentile .1: 248 Percentile 1: 280 Percentile 5: 320 Percentile 10: 350 Percentile 15: 375 Percentile 20: 395 Percentile 25: 410 Percentile 30: 425 Percentile 35: 440 Percentile 40: 455 Percentile 45: 470 Percentile 46: 473 Percentile 47: 476 Percentile 48: 479 Percentile 49: 482 Percentile 50: 485 Percentile 51: 490 Percentile 52: 495 Percentile 53: 500 Percentile 54: 505 Percentile 55: 512 Percentile 60: 540 Percentile 65: 570 Percentile 70: 600 Percentile 75: 640 Percentile 80: 680 Percentile 85: 730 Percentile 90: 790 Percentile 95: 900 Percentile 99: 1150 Percentile 99.9: 1600
I assign a 61% chance that the reported cumulative suspected-case count exceeds 445 by June 1, 2026 00:00 UTC.
WHO says it was alerted on May 5, 2026 to a high-mortality illness in Mongbwalu Health Zone, Ituri; INRB confirmed Bundibugyo virus disease in 8 of 13 analyzed samples on May 15; and DRC declared its 17th Ebola outbreak the same day in Rwampara, Mongbwalu, and Bunia health zones (WHO DON602, 16 May 2026). WHO’s formal baseline was 246 suspected cases and 80 deaths as of May 15, with 24 suspected cases in isolation, 65 listed contacts, weak follow-up due to insecurity, and unusual compatible community-death clusters under investigation in other health zones in Ituri and North Kivu (WHO DON602, 16 May 2026).
The freshest concrete suspected-case count I found is higher: AP reported that by Saturday, May 16, Africa CDC had reported 336 suspected cases, 13 confirmed cases, and 87 deaths in Congo, and warned of “active community transmission” (AP, 16 May 2026). I use 336 as the best operational anchor, but not as a hard floor, because WHO’s May 17 PHEIC statement still used 246 suspected cases as its Ituri figure and because the resolution text has a fallback to 246 if no later suspected-case figure is accepted (WHO PHEIC statement, 17 May 2026).
The completed DRC reference class is small and mostly points to contained outbreaks, but it is not directly the same variable: these are final reported Ebola cases, usually confirmed plus probable, while this question resolves on cumulative suspected cases. Across the 16 completed DRC outbreaks before this one, the median final case count is 56.5 and only 4 of 16 reached 246 or more; the exception that matters most is the 2018-2020 North Kivu/Ituri outbreak, which reached 3,470 cases in the same broad conflict-affected region (CDC outbreak history, updated 15 Dec. 2025; WHO DON602, 16 May 2026).
| DRC outbreak | Species | Final reported cases | Deaths | Main source |
|---|---|---|---|---|
| 1976 Yambuku | Zaire | 318 | 280 | CDC outbreak history |
| 1977 Tandala | Zaire | 1 | 1 | CDC outbreak history |
| 1995 Kikwit | Zaire | 315 | 254 | CDC outbreak history |
| 2007 Kasai Occidental | Zaire | 264 | 187 | CDC outbreak history |
| 2008 Kasai Occidental | Zaire | 32 | 15 | CDC outbreak history |
| 2012 Province Orientale | Bundibugyo | 59 | 34 | WHO DON602, 16 May 2026 |
| 2014 Equateur | Zaire | 69 | 49 | CDC outbreak history |
| 2017 Bas Uélé | Zaire | 8 | 4 | CDC outbreak history |
| 2018 Equateur | Zaire | 54 | 33 | CDC outbreak history |
| 2018-2020 North Kivu / Ituri | Zaire | 3,470 | 2,287 | CDC outbreak history |
| 2020 Equateur | Zaire | 130 | 55 | CDC outbreak history |
| 2021 North Kivu | Zaire | 12 | 6 | CDC outbreak history |
| 2021 Beni | Zaire | 11 | 9 | CDC outbreak history |
| 2022 Equateur | Zaire | 5 | 5 | CDC outbreak history |
| 2022 Beni | Zaire | 1 | 1 | CDC outbreak history |
| 2025 Kasai | Zaire | 64 | 45 | WHO DON589, 1 Dec. 2025 |
The better short-horizon analogues are early suspected-count trajectories. In the 2007 Uganda Bundibugyo outbreak, WHO reported 51 suspected cases on November 30 and 93 suspected cases on December 7, a 1.82x rise in nine days; CDC later lists 131 final reported cases for that outbreak (WHO, 30 Nov. 2007; WHO, 7 Dec. 2007; CDC outbreak history). In the 2012 DRC Bundibugyo outbreak, WHO reported 28 total cases as of September 3, including 14 suspected, and the final total was 59 confirmed or probable cases, a little over 2x the early count (WHO, 5 Sept. 2012; WHO DON602, 16 May 2026). In the 2025 DRC Kasai outbreak, the first public report had 28 suspected cases on September 4 and the final outbreak total was 64 confirmed or probable cases on December 1 (WHO AFRO, 4 Sept. 2025; WHO DON589, 1 Dec. 2025).
The threshold math is the strongest single reason to be above 50%. From the AP/Africa CDC anchor of 336, the count needs 110 more suspected cases to exceed 445, or about 7.3 additional reported suspected cases per day over roughly 15 days (AP, 16 May 2026). That is modest for a PHEIC-scale event where WHO says surveillance is being strengthened, contact listing is immature, and compatible community-death clusters are still under investigation (WHO DON602, 16 May 2026). It is also much smaller than the one-day jump from 246 to 336 in Africa CDC/AP reporting, though I treat that jump as partly case-finding and data reconciliation rather than pure new transmission (AP, 16 May 2026).
The downside case is also real. WHO’s formal count still says 246, and suspected-case categories can be cleaned up as lab results, deduplication, and line-list harmonization catch up (WHO DON602, 16 May 2026; Africa CDC, 15 May 2026). I therefore modeled the resolution as a mixture: 2.5% on exactly 246, 1.0% on 247-335, 5.0% on exactly 336, 30.5% on 337-445, and 61.0% above 445.
The obvious anchor is 246. I think that is stale. AP’s Africa CDC-sourced 336 figure makes the open-ended “above 445” bin much closer than it looks, because the count needs only a one-third rise by June 1 (AP, 16 May 2026). That update is not as clean as a WHO line list, so I do not treat it as certain. But the resolution accepts reputable news outlets, so it is decision-relevant.
The other trap is treating suspected cases as confirmed infections. Suspected cases are a surveillance product. They can rise because teams search harder, investigate deaths, and apply a broad clinical definition in an area with malaria and other febrile illness; they can also stop rising if official reports switch to confirmed and probable cases (WHO DON602, 16 May 2026). For this question, stronger surveillance can increase the resolving number even while control improves.
The main gap is that I could not verify a public Africa CDC bulletin or dashboard carrying the 336 suspected-case figure; the clean public primary-source count remains WHO’s 246, while the higher count comes through AP reporting of an Africa CDC Saturday briefing (AP, 16 May 2026; WHO DON602, 16 May 2026). The second gap is definitional. Later sources may report “suspected,” “suspected plus confirmed,” “alerts,” or only confirmed/probable cases, and the resolver may have to choose among non-identical quantities. The third gap is reporting cadence: if no later suspected-case total is published before June 1, the resolution may sit on a stale figure even if the true outbreak has moved.
Key Update: Africa CDC reported 336 suspected cases as of May 16, 2026—a significant jump from the previously known 246 cases. This new baseline substantially shifts all projections upward.
Baseline reset: The 336-case figure (confirmed by multiple news sources) replaces the outdated 246-case anchor. With 15 days remaining until June 1, forecasts now center around 490–520 total cases rather than the lower estimates based on stale data.
Surveillance catch-up confirmed: The ~90-case jump in one day reflects both new transmission and backlog clearing from cluster investigations, validating expectations that official counts would rise sharply once investigations expanded. This pattern is likely to continue as additional clusters in North Kivu and other Ituri zones are processed.
Growth rate assumptions: Forward projections blend ongoing transmission (estimated 11.7/day historically, potentially higher post-PHEIC) with continued surveillance catch-up. Conservative scenarios assume 4–6 new cases daily; moderate scenarios assume 8–10; acceleration scenarios assume 15–20 daily.
Upper bound now more likely: The previous upper bound of 445.5 cases has roughly 70–75% probability of being exceeded given the 336 baseline, reflecting genuine outbreak size rather than reporting delays.
Uncertainty remains on post-PHEIC trajectory: The May 16 figure predates the May 17 PHEIC declaration, so the impact of heightened response efforts on transmission rates is not yet observable in the data.
Forecast (percentiles, aggregated):
Per-model forecasts:
Rationale: Summary We estimate a median of 600 suspected cases by June 1, 2026, with a 10th to 90th percentile range of 420 to 1050. This reflects an expectation of steady accumulation over the next 15 days, driven by intense contact tracing and surveillance. The outbreak in the Democratic Republic of the Congo’s Ituri Province involves the Bundibugyo ebolavirus strain, for which there are no approved vaccines or treatments. Containment relies entirely on traditional, labor-intensive public health measures. The reported suspected case count has already demonstrated rapid growth, jumping from 246 on May 15 to 336 on May 16. Following the May 17 Public Health Emergency of International Concern (PHEIC) declaration, an influx of international resources will likely trigger aggressive retroactive case finding. The complex operating environment, characterized by mining-related mobility, proximity to international borders, and ongoing insecurity, further elevates the risk of widespread transmission and massive contact lists.
Strongest Arguments for Higher Values
Strongest Arguments for Lower Values
Key Uncertainties
The best current baseline is already messy. WHO’s Disease Outbreak News dated May 16, 2026 says that, as of May 15, Ituri had 246 suspected cases and 80 deaths across Rwampara, Mongbwalu, and Bunia; 24 suspected cases were in isolation, only 65 contacts had been listed, follow-up was weak, and additional compatible clusters were under investigation elsewhere in Ituri and North Kivu. WHO also notes that the first currently known suspected case had symptom onset on April 24, 2026, implying a multi-week detection gap before confirmation on May 15. That combination usually makes retrospective upward revisions more likely. (who.int)
At the same time, reputable reporting had already moved higher by May 16-17. AP’s May 16 reporting from an Africa CDC briefing said that by Saturday, May 16, Congo had 336 suspected and 13 confirmed cases, with 87 deaths, active community transmission, and many active cases still in the community, especially in Mongwalu. On May 17, WHO’s PHEIC statement still cited 246 suspected cases as of May 16 while emphasizing the same broad risk factors. Because the resolution uses the most recently published figure from any reputable outlet, not only official WHO tallies, this reporting conflict is itself part of the forecast. (apnews.com)
The directional forces mostly point upward from here. WHO said the high positivity rate of the initial samples, exported cases to Kampala, clusters of deaths, insecurity, mobility, and informal healthcare networks all suggest a potentially larger outbreak than currently detected. Africa CDC and AP reporting likewise indicate that active community cases, mining-related movement, and insecurity are complicating containment and contact tracing. Those are conditions under which suspect-case counts often rise faster than confirmed-case counts early in an outbreak, especially when line lists are still being reconciled. (apnews.com)
The main moderating factor is response capacity. DRC has extensive Ebola experience, and by May 15-17 WHO, Africa CDC, and MSF were already scaling up surveillance, isolation, lab support, logistics, infection-control work, and community engagement. So I do not think the initial jump from 246 to 336 should be extrapolated linearly for two full weeks. That is an inference from the response actions and DRC’s past performance rather than a directly reported forecast. (afro.who.int)
For base rates, recent DRC outbreaks usually added cases over the first couple of weeks, but the increments were more often in the tens than in the hundreds once the initial retrospective backlog was identified: the 2025 Kasai outbreak started with 28 suspected cases on September 4, 2025 and reached 64 total cases by September 28, 2025; the 2018 Equateur outbreak went from 32 total cases on May 9, 2018 to 58 on May 21, 2018. Earlier Bundibugyo outbreaks were much smaller at notification: WHO recorded 51 suspected cases in Uganda on November 28, 2007 and 10 suspected cases in DRC on August 17, 2012. These analogues are imperfect because they mix suspected, probable, and confirmed definitions differently, but they argue against assuming that the June 1 figure must automatically explode far beyond the current low-to-mid 300s. (afro.who.int)
So my forecast is a mixture. I leave a small exact-mass chance on 246 and 336 because of current publication ambiguity and the possibility that no clearer suspect-case total supersedes them. The main mass is centered in the mid-300s, reflecting continued backfill, line-list harmonization, and additional suspected-case finding over the next two weeks. I also keep a meaningful upper tail in case additional linked zones, backlog deaths, or broader geographic spread are rolled into the tally before June 1, 2026 00:00 UTC. The underlying mixture is centered around about 360 suspected cases, with the densest region roughly in the 330-390 range and a low-double-digit chance that the final published figure exceeds 445. This final paragraph is my inference from the sources above, not a reported estimate. (who.int)
The key forecasting issue is that the question’s fallback figure of 246 already looks stale. WHO’s 16 May Disease Outbreak News said that as of 15 May there were 246 suspected cases, 80 deaths, eight lab-confirmed cases, and 24 suspected cases in isolation across Bunia, Rwampara, and Mongbwalu, with the first currently known suspected case dating back to 24 April. WHO also said only 65 contacts had been listed and that follow-up was weak. But AP, citing Africa CDC’s Saturday briefing, reported 336 suspected cases and 13 confirmed cases by 16-17 May. Africa CDC’s own 15 May release stressed that its figures were provisional and still being validated through laboratory confirmation, line-list harmonization, and epidemiological investigation. (who.int)
The forces pushing the total higher are strong. WHO’s 17 May PHEIC statement said there are significant uncertainties about the true number of infected people and geographic spread, that eight of the first 13 samples were positive, and that increasing syndromic reporting plus clusters of deaths point to a potentially much larger outbreak than currently detected. WHO and Africa CDC also emphasize the weak infection-control environment, deaths among healthcare workers, insecurity, mining-related mobility in Mongbwalu, intense population movement, and urban or semi-urban spread around Bunia and Rwampara. Bundibugyo virus has no licensed vaccine or specific approved therapeutics. (who.int)
The main counterweights are response and time horizon. WHO says rapid response teams are deployed, surveillance and contact tracing are ongoing, laboratory capacity is being strengthened, treatment centers are being set up, and extra supplies and experts are being sent to Bunia; WHO separately said five metric tons of supplies were being airlifted there. Historical analogs show that early Ebola counts usually rise after detection because retrospective investigation finds older cases, but they do not necessarily explode within two weeks. In the 2012 DRC Bundibugyo outbreak, WHO went from 10 suspected cases on 17 August to 41 confirmed/probable cases plus 28 suspected under investigation by 12 September after retrospective work. In Uganda’s 2022 Sudan outbreak, WHO reported 36 confirmed/probable cases by 25 September, and Africa CDC reported 43 by 27 September. (who.int)
So I modeled the resolution as a scenario mixture rather than a single epidemic curve: a small data-harmonization/stagnation scenario, a larger controlled-growth scenario, a sustained backfill-and-expansion scenario, and a right-tail scenario where the current outbreak is materially larger than the line list now shows. Because the resolution target is suspected cases rather than confirmed cases, surveillance expansion can move the count faster than lab confirmations. My distribution is centered a little above 400 suspected cases, with most mass in the high 300s through low 430s and a meaningful but not dominant tail above 445. I leave some probability below 336 because the most recent figures are provisional and undergoing harmonization, but I give very little weight to 246 because later reputable reporting has likely superseded it. (africacdc.org)
As of Sunday, May 17, 2026, the cleanest official WHO baseline is still 246 suspected cases in Ituri as of May 16, alongside 8 laboratory-confirmed cases and 80 suspected deaths. WHO says the first currently known suspected case had symptom onset on April 24; 24 suspected cases were in isolation on May 15; only 65 contacts had been listed, with follow-up weakened by insecurity; and unusual compatible clusters were already being investigated beyond the three named health zones, including elsewhere in Ituri and North Kivu. (who.int)
But the public reporting stream is already inconsistent. AP and other outlets quoting Africa CDCs May 16 briefing reported 336 suspected cases and 13 confirmed cases, while another reputable summary of the same briefing described 336 as the total including suspected and confirmed cases. WHO also corrected its own same-day wording on a Kinshasa case from confirmed to negative on confirmatory testing. Because this market resolves to the latest published suspected-case figure, not the true epidemiological burden, this reporting noise is material. (apnews.com)
I still lean upward from the low 300s because the outbreak has several signatures of under-ascertainment at discovery. Africa CDC described active community transmission and many active cases still in the community; WHO and Africa CDC highlighted mining-related mobility, healthcare-worker infections, insecurity, weak contact follow-up, and significant uncertainty about the true size and geographic spread. A recent DRC comparison also points upward: during the September 2025 Kasai outbreak, suspected cases rose from 28 to 68 within about a week after declaration. (apnews.com)
The main reason I do not push the forecast much higher is that the response is mobilizing hard and the horizon is short. WHO said rapid response teams were deployed, 5 metric tonnes of supplies were being airlifted to Bunia, and a May 17 coordination meeting brought together more than 220 participants to reinforce surveillance, laboratory work, clinical care, logistics and cross-border readiness. Historical Bundibugyo outbreaks also provide a soft brake: WHOs outbreak chronology lists 149 cases in Uganda in 2007 and 57 in DRC in 2012. Those older totals are not directly comparable to todays suspected-case metric, but they argue against an extremely fat upper tail over only the next two weeks. (afro.who.int)
My distribution therefore uses three scenarios: 18% reconciliation or slow growth, 58% moderate growth, and 24% larger hidden-burden growth. I allow the starting point to sit around the official 246, a low-320s interpretation if the Africa CDC 336 figure included confirmed cases, or the mid-330s to 350 range if the higher public count is the correct suspected-case baseline. I then add overdispersed additional reports over the roughly 14-day window to June 1. This yields an implied mean near 394, centers the mass in the high 300s to low 400s, leaves meaningful left-tail probability in the low 300s because of reporting reconciliation risk, and keeps a meaningful but minority tail above 445 because WHO itself says the outbreak may be substantially larger than currently detected and reported. This last paragraph is my inference from the cited evidence. (who.int)
I am forecasting the latest exact suspected-case figure that a reputable source will publish before 2026-06-01 00:00 UTC. The main complication is that reporting is already inconsistent: WHO’s 16 May Disease Outbreak News and its 17 May PHEIC statement still give Ituri 246 suspected cases and 80 suspected deaths, while AP’s report from the 16 May Africa CDC briefing says the count had already risen to 336 suspected and 13 confirmed cases, with 87 deaths. Because this market resolves on the most recently published suspected-case figure, not on the true underlying number of infections, that discrepancy matters a lot. (who.int)
On underlying outbreak dynamics, the pressures are mostly upward. WHO says the first currently known suspected case had symptom onset on 24 April 2026, so confirmation on 15 May came after a multi-week detection gap. WHO and Africa CDC also describe unusual clusters of community deaths, suspected cases across Ituri and North Kivu, weak contact follow-up, high initial sample positivity, insecurity, mining-related mobility in Mongbwalu, and cross-border spread to Uganda. WHO explicitly says these factors point to a potentially much larger outbreak than is currently being detected and reported. (who.int)
There are still real short-horizon constraints on how far the resolving figure can move by June 1. DRC authorities, WHO, Africa CDC and MSF have already deployed rapid response teams, isolation facilities, surveillance, laboratory support, infection-prevention measures, community engagement, and cross-border coordination. Over a two-week horizon, stronger case isolation and field response can slow growth even if retrospective case-finding continues. (who.int)
A second key input is reporting behavior. In the 2025 DRC Ebola outbreak, WHO initially published 28 suspected cases on 5 September, but later official situation reports shifted to confirmed/probable totals rather than maintaining a running suspected-case series. That historical pattern makes sticky early suspected counts more plausible here than a pure transmission model would suggest; this question has genuine reporting-process risk, not just epidemiological risk. (who.int)
So I used a mixture model. I put 5% on an exact 246 outcome, representing a world where no later exact suspected-count figure overtakes WHO’s 17 May publication; 5% on an exact 336 outcome, representing a world where the Africa CDC/AP figure becomes the last exact suspected-count figure repeated later; 20% on an official-update track starting from 246 with modest additional growth; 43% on a moderate case-finding plus ongoing-transmission track starting from 336 with mean additional growth of 80 suspected cases; and 27% on a larger hidden-burden track starting from 336 with mean additional growth of 160. The growth tracks use negative-binomial increments to allow for overdispersion from delayed investigations, clustered community deaths, and line-list reconciliation. This structure is my inference from the current evidence, not a directly reported forecast. (who.int)
That mixture gives an expected value of about 400 suspected cases and roughly 30% probability that the resolving figure ends up above 445. My central mass is in the high-300s to low-400s, but I keep visible point mass on 246 and 336 because exact suspected-case reporting may freeze or become irregular before June 1. (who.int)
I think the key issue is that this market is about reported suspected cases, not confirmed infections. WHO’s DON602 and WHO Africa material still anchor the outbreak at 246 suspected cases, 80 suspected deaths and 8 laboratory-confirmed cases in Ituri across Mongbwalu, Rwampara and Bunia; WHO also says the earliest known suspected case had symptom onset on April 24, WHO was only alerted on May 5, 24 suspected cases were in isolation, 65 contacts had been listed, follow-up was weak, and compatible death clusters were being investigated in other health zones and North Kivu. That combination implies both real transmission and a meaningful retrospective reporting backlog. (who.int)
A second, equally important fact is that later reputable reporting was already higher than 246. AP reported that by the Saturday Africa CDC briefing the tally had reached 336 suspected and 13 confirmed cases, with 87 deaths, and ABC separately reported the same 336 suspected / 13 confirmed figure. But Africa CDC’s own 15 May statement stressed that the numbers were provisional and still being validated through laboratory confirmation, line-list harmonization, contact identification and epidemiological investigation. WHO’s DON602 also shows that some details were still being revised in real time: it says the Kinshasa returnee later tested negative on confirmatory testing, despite some contemporaneous media references to a Kinshasa confirmed case. So the dominant uncertainty here is not only epidemic growth, but also reporting-definition and reconciliation noise. (apnews.com)
For the upside tail, the outbreak has several bad features. WHO and Africa CDC describe an origin in a high-traffic mining area, spread across at least three health zones, intense population movement, insecurity that hampers surveillance and contact tracing, infections among healthcare workers, and cross-border mobility toward Uganda and South Sudan. AP also reported Africa CDC’s statement that 50 deaths had already occurred by the time authorities were first alerted on May 5, which suggests a large amount of retrospective case-finding may still be ahead. On the other hand, WHO and partners are already scaling response: rapid response teams, airlifted supplies, stronger surveillance, lab support, isolation, IPC, and cross-border preparedness were all being reinforced by May 15-17. Early suspected tallies in DRC outbreaks are often noisy and then reporting shifts toward confirmed/probable totals; the 2025 DRC outbreak began with 28 suspected cases reported at declaration and ended with 64 total cases. And CDC’s outbreak history shows the two previous Bundibugyo outbreaks were much smaller in confirmed/final size: 131 cases and 42 deaths in Uganda in 2007, and 38 confirmed cases with 13 deaths in DRC in 2012. (who.int)
My forecast therefore uses a mixture distribution rather than a single extrapolation. I give 10% exact mass to 336, reflecting the chance that 336 remains the last broadly cited suspected-case number before June 1. I then layer a 15% low/reconciled scenario centered in the low 300s, a 50% central scenario centered just under 400, and a 25% high-expansion scenario centered in the mid-450s. That puts the distribution’s center of mass around 395 suspected cases by June 1, 2026 00:00 UTC. I still keep most probability below 445 because only about two weeks remain, because response capacity is ramping up, and because the currently reported deaths-to-suspected ratios of roughly 26%-33% are not wildly inconsistent with WHO’s historical Bundibugyo CFR range of 30%-50%; that argues for continued growth in the reported suspected tally, but not for treating a near-term jump far beyond 500 as the base case. (who.int)