As of Friday, May 8, 2026, the best official anchor is 8 total reported cases. WHO’s May 4 Disease Outbreak News described 7 cases linked to the cruise ship, with illness onset between April 6 and April 28. WHO’s May 7 media note then said 8 cases had been reported so far, including 3 deaths, and that 5 of the 8 had been confirmed as hantavirus. WHO also identified the agent as Andes virus and said more cases might still be reported, while still assessing the overall public-health risk as low. (who.int)

The near-term operational picture points in two directions at once: ongoing case finding, but not obvious runaway spread. Cabo Verde said on May 5 that the 3 patients still on board were clinically stable and that no new incident had been recorded among the other occupants; on May 6 Cabo Verde said those 3 patients had been evacuated. The Netherlands reported 3 symptomatic airplane contacts under testing on May 7, but 2 were already negative and the third was still being analysed. South Africa’s parliamentary briefing said 62 possible contacts had been identified and 42 traced. (minsaude.gov.cv)

Base rates argue against a huge outbreak, but not against a noticeable second wave of linked detections. Andes virus is the only hantavirus with proven person-to-person transmission, usually after close and prolonged contact and mainly during the prodromal phase. In a prospective Chile household study, 16 of 476 household contacts of 76 index cases developed HCPS, or 3.4% overall, with much higher risk for sex partners than for other household contacts. A Chile 2011 cluster produced 5 total cases, including 2 healthcare workers, and reported secondary onsets typically 12 to 27 days after exposure. On the other hand, the 2018-2019 Epuyén outbreak in Argentina reached 34 confirmed infections and 11 deaths, with an estimated median reproductive number of 2.12 before control measures and 0.96 after isolation and quarantine measures were imposed. (gov.uk)

Timing matters a lot for this question. Published incubation estimates for Andes-virus HCPS are broadly about 7 to 39 days after environmental exposure, and CDC and Chile outbreak work suggests person-to-person transmission usually shows symptom onset roughly 9 to 40 days after exposure, often around 12 to 27 days. Because the ship was isolated in early May and exposed contacts are already under tracing across several countries, almost all truly linked symptomatic cases should be identifiable well before the August 1, 2026 resolution date. The main uncertainty is therefore not whether the window is long enough, but how many additional linked cases will be found, and whether some of the currently counted suspected cases are later ruled out. (wwwnc.cdc.gov)

My synthesis is therefore: the mode should stay close to the currently reported count, but the distribution needs a meaningful right tail. I allow some probability that 1 to 3 of the currently counted suspected cases are eventually ruled out, because WHO’s latest split is 5 confirmed out of 8 total. Concretely, in the code I use 72%, 18%, 8%, and 2% on 0, 1, 2, and 3 downward revisions of the current count. For future additions I use a 62%, 25%, 10%, and 3% mixture of contained, moderate, larger, and rare-large spread or case-finding scenarios. This reflects four facts from the sources: authorities have already isolated the ship and evacuated the main symptomatic patients; Cabo Verde reported no new onboard incidents among the remaining occupants as of May 5; WHO explicitly warned that more cases may still be reported; and diagnostic/testing capacity has been expanded across multiple countries. (who.int)

On that basis, I center the forecast around roughly 10 to 11 total publicly reported linked cases by August 1, 2026. My modal region is 8 to 9, most of my probability mass is in roughly the 8 to 15 range, and I keep a long but fairly thin right tail into the 20s and beyond because the MV Hondius involved prolonged close contact in a closed setting and Andes virus has produced multi-step transmission chains before. The biggest argument against a very large number is that control measures and contact tracing started early relative to the observed cluster size; the biggest argument against simply freezing the count at 8 is that WHO and national authorities are still actively finding and testing linked contacts across countries. (who.int)

As of May 7, 2026, WHO had moved the MV Hondius cluster from 7 reported cases on May 4 to 8 reported cases, including 3 deaths, and said 5 of the 8 were confirmed hantavirus cases. WHO identified the agent as Andes virus, noted that limited human-to-human transmission can occur after close and prolonged contact, said more cases might still be reported because of the incubation period, and said it had shipped 2,500 diagnostic kits from Argentina to laboratories in five countries. Spain and Singapore were still describing the same 8 linked cases with fewer cases already laboratory-confirmed and others still under investigation, so the status of several current cases is still somewhat fluid rather than fully settled. (who.int)

The strongest reason not to forecast a very large final number is that the outbreak does not look explosive in real time. Cabo Verde said on May 5 that there had been no new incidents among other occupants, and Spain said on May 7 that the 147 passengers and crew sailing to the Canary Islands were asymptomatic. WHO, RIVM, Switzerland’s FOPH, and UKHSA all publicly assessed the risk to the broader public as low or very low, while national authorities were already using isolation, evacuation, onboard assessment, and contact tracing. Those facts do not rule out more cases, but they argue against an uncontrolled shipwide outbreak. (minsaude.gov.cv)

I still think additional cases are more likely than not. Andes virus is the only hantavirus with documented limited person-to-person spread; WHO says symptoms can begin 1 to 8 weeks after exposure, CDC and Chilean guidance use a 42-day monitoring window, and a Chile cluster paper says symptom onset in person-to-person transmission has usually occurred 12 to 27 days after close contact. That leaves enough time for people exposed in late April or early May to be publicly reported before August 1. There is also active international case finding: RIVM reported three symptomatic airplane contacts tested in the Netherlands, with two negatives and one still pending on May 7; Singapore had two returnees isolated and being tested, one asymptomatic and one with only a runny nose; and UKHSA said it was following a small number of close contacts. (who.int)

For base rates, ordinary close-contact spread of Andes virus has usually been limited rather than highly efficient: a Chile prospective household-contact study found 16 HCPS cases among 476 household contacts, or 3.4%, with much higher risk among sex partners than among other household contacts. But there is a real heavy-tail possibility: the 2018-2019 Epuyén outbreak in Argentina produced 34 confirmed infections and 11 deaths, showing that sustained multi-generation spread can occur in unusual circumstances. WHO also said the first two cases had traveled through Argentina, Chile, and Uruguay before boarding, and Argentina’s epidemiological bulletin shows ongoing hantavirus circulation nationally, so a small number of additional linked cases from shared pre-boarding exposure or late contact tracing cannot be dismissed. (academic.oup.com)

My model therefore starts from the 5 already-confirmed WHO cases, treats the 3 currently suspected cases as likely but not certain to remain in the final deduplicated count, and then adds a mixture distribution for newly detected cases. Most weight is on contained scenarios with only a few more reports, but I keep a meaningful tail for moderate cluster expansion and a small tail for an Epuyén-like surprise. This puts my forecast center at about 11 total cases by August 1, with most of the probability mass in the 8 to 15 range and a thin tail into the 20s and 30s. (who.int)

As of May 7, 2026, WHO said eight cases linked to the MV Hondius cluster had been reported, including three deaths; five of the eight were laboratory confirmed and three were still suspected. In Tedros’s May 7 briefing, WHO also laid out the current eight-case picture: three deaths, one patient still in intensive care in South Africa, three more patients evacuated from the ship, and one Swiss case confirmed after disembarking in Saint Helena; WHO added that none of the remaining passengers or crew still on the ship were symptomatic at that point. WHO also said it is aware of other symptomatic contacts, that Andes-virus incubation can run up to six weeks, that it informed 12 countries whose nationals disembarked in Saint Helena, and that South Africa is following up the flight contacts of the woman who died after flying from Saint Helena to Johannesburg. (who.int)

The near-term follow-up data argue for some additional linked cases but not an explosive international cascade. In the Netherlands, RIVM said on May 7 that three people with symptoms after airplane contact had been tested; two were negative and one was still pending. Singapore said on May 7 that two returnees from the ship were isolated and being tested, with one asymptomatic and one only mildly symptomatic, and that exposed persons could be monitored for up to 45 days. UKHSA said on May 6 that two people had already returned independently to the UK without symptoms and that a small number of close contacts were also asymptomatic. Switzerland confirmed one positive case on May 6 and said the patient’s wife was self-isolating without symptoms. (rivm.nl)

The biological and historical base rate is mixed. WHO and ECDC both emphasize that Andes virus is the only hantavirus known for limited person-to-person transmission, and that transmission has mainly been associated with close and prolonged contact. ECDC says everyone on the ship is being treated as a close contact because of the closed setting, but still assesses the risk to the general EU/EEA population as very low. For Andes virus, published estimates put the median incubation period around 18 days with observed ranges extending from about 7 to 39 days, while Chilean outbreak guidance recommends monitoring close contacts for the full documented 42-day incubation period. The main modern analogue is the 2018-2019 Epuyén outbreak in Argentina: after a single zoonotic introduction, person-to-person spread produced 34 confirmed infections and 11 deaths, with an estimated median reproductive number of 2.12 before control measures and 0.96 after isolation and self-quarantine were imposed. (ecdc.europa.eu)

My synthesis is that the MV Hondius cluster is already too large, and too geographically dispersed, for 8 to be a safe point forecast, but it still looks much more like a contained multi-country tracing event than a replay of Epuyén. The most important reasons are: the shipboard outbreak was recognized and aggressively managed by early May; WHO says the remaining people on board were asymptomatic as of May 7; multiple countries are already isolating, testing, and tracing exposed travelers; and the first signals from off-ship follow-up are mostly negatives or monitored contacts rather than a wave of confirmed secondary cases. Against that, WHO explicitly expects that more cases may be reported, the ship held roughly 150 people in a confined environment, one disembarked traveler in Switzerland has already converted into a confirmed case, and enhanced testing could still uncover mild, suspected, or asymptomatic linked infections that would count under the question rules. Because the main exposure window appears to close in early May and WHO gives a maximum incubation of about six weeks, the August 1, 2026 deadline should capture nearly all primary cases from this event and a large share of any immediate secondary chain. (who.int)

My central forecast is 11 total publicly reported linked cases by August 1, 2026. I put most of the probability mass in the 8 to 15 range, with a meaningful but minority tail into the 20s and a small tail above 40. I also leave a small probability on 5 to 7 because some currently suspected cases could be ruled out before resolution under the question rules.

WHO is the highest-priority source under the resolution rules, so I anchor on its latest update. WHO first reported 7 linked cases on May 4, 2026, then updated that to 8 linked cases on May 7, with 5 confirmed hantavirus infections and 3 deaths. WHO identified the virus as Andes virus, assessed wider public-health risk as low, warned that more cases could still appear because incubation can reach six weeks, and said it had shipped 2,500 diagnostic kits from Argentina to laboratories in five countries. In the Director-General’s remarks the same day, WHO also said the first known case developed symptoms on April 6, 2026, that nationals from 12 countries had already disembarked in Saint Helena, and that the remaining passengers were being medically assessed on board. Because Spain’s May 6 breakdown was both earlier and lower-priority, I treat WHO’s May 7 statement of 8 linked cases and 5 confirmed as the operative starting point. (who.int)

The immediate upside risk comes from follow-up testing and delayed onset among already exposed contacts, but official reports so far do not look like explosive onward spread. Spain said on May 6 that the three symptomatic people still on the ship had already been evacuated and only asymptomatic passengers and crew remained on board, with the rest to move under controlled quarantine/disembarkation procedures. Switzerland publicly reported one cruise-linked positive case on May 6. The Netherlands said on May 7 that three people who developed symptoms after airplane contact with an infected passenger had been tested, with two negative and one still pending. Singapore said on May 7 that two residents who had been on the ship were isolated and tested, with results pending. That pattern implies several plausible additions to the public tally, but also shows that at least some first-wave suspect contacts are already testing negative. (sanidad.gob.es)

For base rates, Andes-virus person-to-person spread is rare but real, and heavily concentrated in very close contacts. In a prospective Chile study of 476 household contacts of index patients, 16 contacts, or 3.4%, developed HCPS; risk was 17.6% among sex partners but only 1.2% among other household contacts. CDC’s Delaware investigation of an imported Andes-virus case traced 53 contacts and found no secondary cases. The downside tail is still meaningful, though: the 2018-2019 Epuyén outbreak in Argentina produced 34 confirmed infections and 11 deaths, and reconstruction of that outbreak found transmission was driven by 3 symptomatic people attending crowded social events, with median R falling from 2.12 before control measures to 0.96 after isolation and quarantine were imposed. A 2011 Chile cluster also produced 4 secondary cases from 1 index patient, including 2 nosocomial infections. (academic.oup.com)

My synthesis is that this cluster is more likely to finish as a medium-sized close-contact outbreak than as an Epuyén-scale event. I infer that because international recognition and control arrived relatively early in the visible outbreak phase: WHO and national authorities are coordinating across countries, Spain said active symptomatic cases were already removed from the ship, and the first publicly reported second-ring testing in the Netherlands already included two negatives. Against that, the six-week incubation window, prolonged shipboard exposure before recognition, WHO’s explicit warning that more cases may be reported, and the fact that one linked case is already asymptomatic all argue against a flat final count of 8. My central forecast is 10 total publicly reported linked cases by August 1, 2026, with most probability on 8-13, a smaller but meaningful tail into the high teens and 20s, and only a very small tail toward outcomes above 30. (who.int)

I anchor the current starting point at 8 linked cases. WHO’s Disease Outbreak News on May 4 reported 7 cases, but Tedros/WHO said on May 7 that 8 cases had been reported, including 3 deaths, with 5 confirmed and 3 suspected. Spain’s May 6 press note still described 8 linked cases but only 3 lab-confirmed, and ECDC’s May 6 brief still summarized 7 cases, so under the stated resolution hierarchy I use the later WHO count and classification as the operative baseline. (who.int)

The virology and transmission pattern argue against a huge outbreak, but not against a modest increase. WHO and ECDC both say Andes virus is the only hantavirus with documented human-to-human spread, and that this spread is limited, uncommon, and associated with close and prolonged contact. ECDC also notes that transmission has mainly been observed in household or intimate-contact settings rather than sustained casual spread, that infectivity is highest around the first day of symptoms, that evidence for asymptomatic transmission is very limited, and that incubation can run from about a week up to six weeks. (who.int)

There are still real reasons to expect some additional linked reports before August 1. WHO has informed 12 countries whose nationals disembarked in Saint Helena, is supporting South Africa in flight-contact follow-up, and has shipped 2,500 diagnostic kits from Argentina to laboratories in five countries. ECDC is treating everyone on board as close contacts as a precaution, and the resolution rules count not just confirmed cases but also suspected/probable cases and explicitly linked asymptomatic infections. That combination makes a small rise from further tracing, retesting, or serology more likely than a flat line. (who.int)

But the control picture is much better than in a worst-case amplification event. Spain said the three symptomatic people had been evacuated and that the 147 people heading to the Canary Islands were asymptomatic on May 7. UKHSA said the British nationals who had independently returned to the UK were asymptomatic, and RIVM said that among 3 symptomatic airplane contacts in the Netherlands, 2 had already tested negative and 1 was still pending. WHO, ECDC, RIVM, the Swiss FOPH, and UKHSA all described the wider public risk as low or very low. (lamoncloa.gob.es)

For base rates, I put weight on two opposite comparators. A CDC investigation after an imported Andes case in the United States identified 53 contacts across six states and found zero secondary cases, which supports low onward transmission when modern contact tracing is in place. The main tail-risk comparator is the 2018-2019 Chubut outbreak in Argentina, which produced 34 confirmed infections and 11 deaths, showing that superspreading is possible even if uncommon. (cdc.gov)

My quantitative model therefore does three things. First, it treats the current 5 WHO-confirmed cases as fixed. Second, it treats the current 3 suspected cases as highly likely, but not certain, to still count by August 1 if not later ruled out. Third, it adds a mixture distribution for additional linked cases: a dominant low-spread scenario, a meaningful moderate scenario driven by case finding and a few incubating infections, and a small superspread tail. That produces an expected total of about 10.82 cases. In words, I think the distribution should be centered in the high single digits to low teens, with a noticeable but smaller shoulder into the teens and 20s, and only a very small chance of anything remotely approaching 100. The key reason I do not go higher is that the ship is already under tight control, the remaining people reported on board were asymptomatic as of May 7, and August 1 is well beyond the six-week monitoring window for May exposures, so most infections seeded by this event should have had time to declare and be reported by then. (who.int)

How many total cases epidemiologically linked to the MV Hondius hantavirus cluster will WHO or relevant national authorities publicly report by August 1, 2026?

• My most likely prediction for total cases epidemiologically linked to the MV Hondius hantavirus cluster is 20.50, with a 50% chance of falling between 14.50 and 28.50.
• As of May 7, 2026, the count has already reached 13 cases (5 confirmed and 8 suspected), establishing a high floor for the final resolution.
• The primary population at risk is the approximately 150 passengers and crew, whose 6-week incubation window remains open through mid-May, with secondary transmission potential extending into June.
• Global contact tracing is intensive, with 2,500 diagnostic kits distributed and monitoring occurring in at least 12 countries, which will likely uncover additional suspected or asymptomatic cases.
• Biological constraints of the Andes virus, which requires close and prolonged contact for limited human-to-human spread, make a large-scale pandemic (100+ cases) highly unlikely based on historical data.
• The arrival of the ship in the Canary Islands on May 10-11 will be a critical juncture for final screening and case reporting.

Key figures

Historical context

• Andes virus (ANDV) is unique among hantaviruses for its documented ability to spread human-to-human.
• The most significant historical precedent is the 2018-2019 outbreak in Epuyén, Argentina, which began with a single rodent-to-human infection and grew to 34 confirmed cases and 11 deaths through person-to-person transmission in social settings.
• Most hantavirus outbreaks are zoonotic (rodent-to-human) and do not result in clusters larger than a single household; cruise ships, however, present a unique high-density environment that mimics the ‘close contact’ required for ANDV spread.
• The incubation period for ANDV is traditionally 8 to 31 days but can extend up to 6 or 8 weeks, meaning cases can appear long after the initial exposure event.

Tailwinds

• Inclusive Case Definition: The resolution includes ‘suspected’ and ‘asymptomatic seropositive’ cases, which are easier to identify through aggressive testing than ‘confirmed’ clinical cases.
• Global Surveillance: With passengers scattered across 12+ countries and 2,500 kits available, cases that might have gone unnoticed in a remote region are likely to be caught and reported.
• Long Incubation: The potential for cases to emerge up to 8 weeks after exposure means the reporting window will remain active through July.
• Ship Environment: The confined nature of the MV Hondius between April 1 and May 10 provided ample opportunity for secondary and tertiary transmission among the 150 individuals on board.

Headwinds

• WHO Risk Assessment: The official ‘low’ risk assessment suggests that international health authorities do not expect efficient or widespread human-to-human spread.
• Biological Limitations: Andes virus typically requires close and prolonged contact for transmission, limiting the potential for third or fourth-generation cases.
• Effective Isolation: High-risk individuals in Singapore, South Africa, and the UK are already in isolation, which should break transmission chains before they expand.
• Case Disqualification: The resolution criteria exclude cases that are initially reported but later ruled out; some ‘suspected’ cases currently in the count may be discarded.

Detailed reasoning

• The forecast is anchored by two primary factors: the current reported case count (which had already reached 13 by May 7) and the finite population of the MV Hondius (approximately 149-150 passengers and crew). Given that 13 cases are already publicly acknowledged by a mix of WHO and national authorities, the floor for this prediction is effectively established.
• I weight the ‘inside view’ (current data) heavily. With 13 cases reported and individuals already under isolation in Singapore, Canada, the US, and the Netherlands, a near-term increase to approximately 18-22 cases is almost certain as pending test results are finalized. The arrival of the ship in Tenerife on May 10-11 will likely trigger a comprehensive screening of the remaining ~110 people on board. If the attack rate on the ship follows historical cruise ship outbreaks, we should expect a small but significant number of additional suspected or asymptomatic infections to be identified during this process.
• I use the ‘outside view’ of historical Andes virus (ANDV) outbreaks as a stabilizing boundary. The 2018-2019 Epuyén outbreak in Argentina is the most relevant precedent. It involved human-to-human transmission in a close-knit community and resulted in 34 cases. While the MV Hondius cluster is more geographically dispersed, the level of global surveillance is significantly higher. The deployment of 2,500 diagnostic kits suggests health authorities are casting a wide net. This aggressive case-finding is likely to identify more “suspected” or “asymptomatic seropositive” cases than would be found in a typical zoonotic outbreak, potentially pushing the count toward the 25-35 range.
• However, the “limited” nature of ANDV human-to-human transmission is a strong counter-weight. Historically, transmission chains rarely exceed three generations. Unlike respiratory viruses like COVID-19 or Influenza, ANDV requires close, prolonged contact (e.g., household members or medical staff). This biological constraint makes an exponential explosion into the hundreds highly unlikely. The WHO’s assessment of the risk as “low” and the lack of evidence of community transmission in South Africa or the Netherlands further support a contained forecast.
• The 50% HDI of 14.50 to 28.50 reflects a scenario where the outbreak is largely confined to the initial passengers and their immediate household contacts. The upper tail of the 80% HDI (up to 41.50) accounts for a scenario similar to Epuyén, where a few secondary transmission chains in the passengers’ home countries lead to small localized clusters. I assign very low probability to values over 100, as this would require a level of transmissibility not previously seen in Andes virus.

Key uncertainties

• Accuracy of ‘Suspected’ Reporting: If national authorities report many individuals as ‘suspected’ based on mild symptoms alone, the count could artificially inflate.
• Asymptomatic Rate: The number of passengers who may test positive for antibodies without ever showing symptoms is unknown but could significantly boost the total if broad serology is conducted.
• Tenerife Screening Results: The outcome of the mass screening of the ~110 remaining people on the ship will determine if the cluster is already peaking or just beginning to grow.
• Transmission Chain Length: Whether the virus can sustain more than 3 generations of transmission in this specific outbreak remains an epidemiological uncertainty.

Conclusion

• My most likely prediction for the total number of epidemiologically linked cases is 20.50, with a 50% chance of falling between 14.50 and 28.50.
• While the initial cruise ship cluster is finite, the inclusion of ‘suspected’ and ‘asymptomatic’ cases in the resolution criteria, combined with aggressive global contact tracing, ensures the final count will likely double or triple from the initial May 4 baseline.
• Decision-makers should focus on results from Tenerife and South African contact tracing as the primary indicators of whether this outbreak will remain in the low double-digits or approach historical peaks of 30-40 cases.

Question: will the final MV Hondius–linked Andes virus case count by Aug 1 exceed the current WHO anchor of 8 epidemiologically linked cases? The authoritative anchor is WHO’s 7 May 2026 count of 8 linked cases (5 confirmed, 3 suspected, including one asymptomatic), and forecasters agree the main remaining upward channels are delayed symptomatic onsets among the finite ship cohort (incubation up to ~42–45 days, so risk through mid‑June), limited human‑to‑human secondary transmission to household/healthcare/flight contacts, and retrospective serology finding asymptomatic infections. Constraining factors noted across teams are that the passenger/crew cohort is well enumerated, several early contact investigations have been negative, and no broad source outbreak has been identified at embarkation, so expansion like the Epuyén community outbreak is considered unlikely though not impossible. Quantitatively one team places modal mass at no additional cases (40% chance) or only 1–3 extra (39%), with smaller probabilities for larger increases (15% for 4–7 extra, 4% for 8–12 extra, ~2% >20); other teams similarly concentrate probability around 8–12 with a thin long tail into the 20s–30s. Possible disagreement is small: some forecasters allow a modest chance the count could drop if suspected cases are ruled out, but overall the consensus is modal stability at ~8 with modest probability of growth into the low teens and a long, thin tail for larger outbreaks.

Forecast and Reasoning

Forecast: Probability distribution:

• 5.00% chance of value below 5.0
• 10.00% chance of value below 6.0
• 20.00% chance of value below 8.0
• 30.00% chance of value below 10.0
• 40.00% chance of value below 13.0
• 50.00% chance of value below 16.0
• 60.00% chance of value below 19.0
• 70.00% chance of value below 23.0
• 80.00% chance of value below 30.0
• 90.00% chance of value below 42.0
• 95.00% chance of value below 58.0

Forecast Reasoning

As of May 7, 2026, WHO public statements put the MV Hondius-associated cluster at 8 reported cases (5 confirmed, 3 suspected). Between now (May 8) and Aug 1 there is enough time to (a) finish follow-up covering the stated long incubation window (up to ~8 weeks) and (b) complete multi-country testing/contact tracing that could add cases, including any explicitly reported asymptomatic seropositives. At the same time, major expansion is constrained by (i) a finite primary exposed cohort (the ship) and (ii) Andes virus human-to-human spread being described as limited and typically requiring close/prolonged contact, plus rapid international control measures.

So the main mass should be in “low double digits to a couple dozen”: some probability the total ends close to today’s count (if suspected cases are later ruled out and few/no additional positives are found), but a meaningful chance of additional ship-linked detections plus a small number of close-contact secondary cases across countries. A much larger total (40–60+) is a tail scenario requiring either unusually extensive secondary transmission or unusually expansive public reporting of mild/asymptomatic infections; >100 by Aug 1 looks very unlikely, so I keep all percentiles below the open upper bound rather than placing mass past 100.

An amateur forecast would commonly either (1) anchor on the current WHO count (~8) and underweight the remaining follow-up/testing window, or (2) over-extrapolate from “cruise ship outbreak” framing and assume near-Diamond-Princess-style attack rates, ignoring Andes virus’s typically limited human-to-human transmission and the strong role of close-contact chains. My forecast explicitly balances (a) scope for additional case-finding and cross-national reporting with (b) biological/epidemiological constraints and known historical cluster sizes, yielding a median modestly above today’s total but with a calibrated right tail into several dozen cases. I’m moderately confident this is better calibrated than either extreme.

Forecast rationale (numeric):

— Iteration 1 — Across the forecasts, the core reasoning is that the publicly reported total is most likely to rise only modestly from the current baseline of 8 cases, with the most plausible outcome landing in the low teens by Aug. 1, 2026.

Main factors driving the outlook

• Current baseline matters most: Both forecasts anchor on the then-public count of 8 total cases and assume this remains the starting point for the projection.
• Ongoing contact monitoring could add cases: Roughly 149 people under observation creates room for additional detections, especially among close contacts.
• Some secondary transmission is possible, but likely limited: The models allow for a small number of household or healthcare-associated secondary cases, but do not expect broad spread.
• Reclassification can move counts both ways: Some currently suspected cases may later be ruled out, partially offsetting any new confirmed links.
• Asymptomatic or serologic links could be counted: A few forecasts explicitly include the possibility that authorities may publicly count linked asymptomatic seropositive infections if detected.
• Rare outbreak expansion is the main tail risk: While judged unlikely, both models keep open the chance of a larger-than-expected cluster if transmission chains extend farther or if reporting definitions broaden.

Areas of consensus

• Direction of travel: More cases than 8 is more likely than not.
• Magnitude: Growth is expected to be modest, not explosive.
• Central estimate: The forecasts cluster around the low teens.
• Uncertainty is asymmetric: Downside is possible via rule-outs, but the models put meaningful weight on upside risk from delayed detection or broader case classification.

Differences in emphasis

• One forecast centers on a median of about 12 and stresses a modest increase with a fat upper tail.
• The other is slightly more optimistic on case growth, with a median of about 13 and an expectation around 14, while also explicitly modeling new primary, secondary, and asymptomatic detections.
• Both retain very wide tails for rare scenarios, but agree those outcomes are very unlikely relative to the low-teens center.

Bottom line

The collective view is that the MV Hondius cluster will probably be reported at around 12–13 total cases by Aug. 1, 2026, driven by a few additional detections among monitored contacts and possibly limited secondary spread, with reclassification and definition changes creating some uncertainty.

— Iteration 2 — The forecasts converge on a modest upward revision from the latest publicly reported total of 8 cases, with the most likely outcome by Aug. 1, 2026 in the low teens.

Main reasoning patterns

• Start from the latest WHO count: Both forecasts anchor on the publicly reported total of 8 cases (5 confirmed, 3 suspected) as the baseline.
• Expect incremental growth, not explosive spread: The cluster is treated as a contained investigation with ongoing contact tracing and testing, so only a small number of additional cases are expected.
• Allow for reclassification and case definition effects: Some additional publicly counted cases could come from:
  • newly confirmed symptomatic cases among exposed contacts,
  • serology-positive or asymptomatic linked cases being explicitly reported,
  • or suspected cases being reclassified downward, which could slightly reduce the total.
• Low transmission risk constrains the forecast: Both rationales emphasize that human-to-human spread appears limited, keeping the central estimate from rising sharply.
• Right-skewed uncertainty: While the center is modest, both forecasts retain a long right tail for low-probability scenarios where investigation expands more than expected, including broader serologic findings across jurisdictions.

Areas of consensus

• Central estimate: roughly 12–14 total cases by Aug. 1.
• Most likely range: about 8–20, with most probability mass in a contained-investigation band.
• Bottom line: a small increase is more likely than a major outbreak expansion.

Differences in emphasis

• One forecast is slightly more bullish on case growth, centering around 14.
• The other is a bit lower, centering around 12.
• Both, however, share the same core view: modest additions are plausible, but the outbreak is unlikely to balloon dramatically without unexpected new evidence.

— Iteration 3 — The forecasts converge on a modest increase from the current WHO baseline of 8 cases (5 confirmed, 3 suspected) rather than a large outbreak.

Main reasoning patterns

• Starting point matters: Both anchor on WHO’s publicly reported 8 cases as of May 7.
• Limited expected growth: The most likely outcome is that some suspected cases are confirmed and a few additional linked cases are found through tracing, onboard/arrival assessments, and multi-country follow-up.
• Offset from reclassification: Some suspected cases may be ruled out, which tempers the total.
• Low-transmission expectation: The reasoning assumes no major sustained human-to-human spread, consistent with the low-risk assessment and typical behavior of hantavirus clusters.
• Conservative uncertainty handling: Both allow for a small chance of much higher totals if broader testing identifies more exposed or asymptomatic cases, or if the investigation uncovers a larger shared-source event than currently known.

Areas of consensus

• The central estimate is in the low double digits.
• A reasonable implied range is roughly 8–20 cases, with the most likely point estimate around 10–12.
• The distribution is right-skewed: modest increases are expected, but rare scenarios could push totals much higher.

Bottom line

Overall, the collective reasoning expects only a limited rise above the initial cluster size, with low double-digit total cases by August 1, 2026 being the most plausible outcome.

Based on the provided rationales, the forecasts for the final publicly reported case count of the MV Hondius hantavirus cluster rely on the following shared arguments and evidence:

Current Baseline and Reporting Timeline Forecasters anchor on the existing baseline of confirmed and suspected cases, noting that the outbreak has already caused fatalities and medical evacuations. Because the maximum incubation period for the Andes orthohantavirus (ANDV) is approximately eight weeks, individuals exposed before the ship’s early May quarantine will develop symptoms by mid-to-late June. Consequently, the August 1 resolution deadline provides ample time for health authorities to process laboratory tests, officially rule out false alarms, and finalize the public count.

Transmission Dynamics and Exposure Windows ANDV is unique among hantaviruses for its human-to-human transmission capability. However, transmission is highly inefficient, requiring close, prolonged contact such as sharing a confined cabin, sexual intimacy, or providing direct healthcare without adequate personal protective equipment. Because of this, forecasters agree that an exponential, airborne cruise-ship outbreak is highly unlikely. Still, because early symptomatic cases mixed freely with other passengers and took international flights before strict maritime isolation was enforced, a moderate “tertiary wave” of new infections among close contacts, flight crews, and medical staff is highly expected.

Testing Breadth and Asymptomatic Surveillance An aggressive, multinational contact-tracing operation is underway across multiple countries. Many rationales highlight that the wide deployment of diagnostic test kits could uncover mild or entirely asymptomatic cases that usually go undetected, modestly inflating the final tally. Conversely, some forecasters note that certain health agencies advise against routinely testing asymptomatic individuals, which may limit the number of “silent” cases added to the official count.

Historical Precedents and Scenarios Forecasters consistently reference historical human-to-human ANDV clusters—specifically the Epuyén and El Bolsón outbreaks in Argentina—to establish the likely ceiling for this event.

• Optimistic Scenario: Strict maritime quarantine perfectly severed transmission chains, and a portion of the currently “suspected” cases test negative, keeping the final count close to the current baseline.
• Central Scenario: The prolonged pre-quarantine window generates a steady trickle of new cases among high-risk contacts (e.g., cabin mates and medical responders), leading to a moderate, localized expansion before the outbreak is extinguished.
• Pessimistic Scenario: The dense, unmitigated mixing prior to the ship’s lockdown acted as a super-spreader environment, mirroring the worst historical ANDV outbreaks and driving a significantly larger wave of tertiary cases.

Agent Forecast (discrete, confidence: medium)

• 13: 10.5%
• 14: 10.5%
• 12: 9.5%
• 15: 9.5%
• 16: 8.5%

Base Rate: Documented Andes virus human-to-human clusters historically range from 3 to 16 cases per event, with the 2018–19 Epuyén event (34 cases via super-spreader social gatherings) as the notable outlier. The Hondius cluster’s 8 qualifying cases as of May 7–8 already sits near the upper end of typical clusters, suggesting modest growth (to ~13–16) is more plausible than either stagnation at 8 or escalation to Epuyén-scale.

Initial Reasoning:

• As of May 7–8, 2026, WHO officially reports 8 cases (5 confirmed, 3 suspected) with 3 deaths aboard the MV Hondius.
• Media sources (ABC News, CBS News) report 9–12+ cases, incorporating French and Swiss contacts, KLM flight attendant, and 3 evacuees from May 6.
• The last illness onset was April 28; the maximum 42-day incubation closes around June 9, beyond which new symptomatic cases are impossible. 2,500 diagnostic kits were shipped to 5 countries for systematic testing.
• The resolution criteria includes asymptomatic seropositive individuals explicitly reported in the cluster investigation, creating a wildcard: if ~5–10% of 147 passengers/crew seroconvert and are officially reported, the total could rise into the 15–25 range.
• Historical Andes virus H2H clusters typically involve 3–16 cases; the largest (Epuyén 2018–19) reached 34 via super-spreader dynamics at social events, which differ substantially from this contained-ship scenario.
• No monitored contacts in 5 US states or 12+ countries have shown symptoms as of May 7–8, supporting containment.
• Mode at 13–14, 80th-percentile ~19, 95th-percentile ~25.

Key Uncertainties:

• Whether national authorities (NL, FR, CH, DE, UK, US) will officially classify currently-pending suspected cases before August 1
• Whether any serological survey data on the 147 passengers/crew is publicly published with explicit cluster attribution
• Whether new symptomatic secondary cases emerge in healthcare or household contacts before the incubation window closes (~June 9)
• Whether WHO issues a consolidated DON update closer to August that aggregates national reports
• Whether the 30 passengers who disembarked at St. Helena before identification produce additional classified cases

Evidence Sources (14):

• WHO Disease Outbreak News DON599 – May 4, 2026
• WHO Director-General briefing – May 7, 2026
• MV Hondius hantavirus outbreak – Wikipedia
• CBS News – Hantavirus Canary Islands cruise ship new cases Switzerland
• CBS News – 5 US states monitoring MV Hondius passengers
• ABC News Facebook post – MV Hondius cases risen to at least 12
• NEJM 2020 – Super-Spreaders and Person-to-Person Transmission of Andes Virus in Argentina
• Chile ANDV household contact study – secondary attack rates
• CDC EID 2014 – Person-to-Person ANDV Transmission Chile 2011
• BBC – Hantavirus-hit cruise ship on way to Canary Islands
• ECDC Risk Assessment – Hantavirus cluster on cruise ship, May 6, 2026
• Dr. Céline Gounder – Hantavirus Risk, Prevention, Treatment (May 6, 2026)
• TODAY.com – Will the Hantavirus Cruise Ship Cause a New Pandemic? (May 7, 2026)
• Snopes Fact Check – Hantavirus human-to-human transmission history

Grounding Facts (passed to reviewers):

1. As of May 7-8, 2026, the number of suspected or confirmed hantavirus cases associated with the MV Hondius cruise ship has risen to at least 12.
2. As of May 7, 2026, there are 3 reported deaths linked to the MV Hondius hantavirus cluster.
3. The MV Hondius carries 147 people (88 passengers, 59 crew) of 23 nationalities.
4. Illness onset dates for reported cases were between April 6 and April 28, 2026.
5. The MV Hondius sailed to Tenerife on May 6, 2026, and infected passengers have been hospitalized in South Africa, Netherlands, Germany, and Switzerland.
6. As of May 7-8, 2026, at least 12 countries and 5 US states are monitoring individuals from the MV Hondius for hantavirus.
7. WHO Director-General briefed media on May 7, 2026, stating 8 cases (5 confirmed) and maintaining a “low” risk assessment for the outbreak.
8. The deadline for publicly reported cases for the question is August 1, 2026.
9. Past Andes virus outbreaks have demonstrated person-to-person transmission, including household and healthcare worker transmission, with a 2018–2019 Argentine outbreak resulting in 34 confirmed infections and 11 deaths over 15 weeks.

Draft: 13: 10.0%, 14: 10.0%, 12: 9.0%, 15: 9.0%, 16: 8.0%

Reviewer Panel:

Reviewer (reasoning)

The draft’s reliance on a speculative 5–10% asymptomatic seroconversion rate among all passengers to drive the 15–25 case tail is the weakest link, since public reporting of such serosurveys is uncommon and no preliminary serological data have been released. It also overstates the certainty that incubation ends by June 9, ignoring potential secondary transmission in healthcare settings that could yield onset beyond that date. Finally, the forecast’s detailed probability mass function lacks a clear rationale for its long upper tail (e.g. 19–25 cases) given containment measures and current case counts.

Objections:

• Assumes large asymptomatic seropositive case counts (5–10% of 147) will be officially reported, despite no indication WHO or national agencies plan to publish full serosurvey data.
• Treats 42-day incubation as a hard cutoff, ignoring real-world delays in detecting hospital-acquired secondary cases, which could push onset past early June.
• Attributes nonzero probability to very high totals (e.g. >30) without justifying a plausible transmission chain or failure of containment beyond the cruise ship.

Reviewer (calibration)

The draft forecast’s baseline count of 8 cases from WHO contradicts the verified grounding fact of “at least 12” cases, leading to an artificially low starting point. It also appears to underplay the potential for secondary human-to-human transmission by dismissing the Epuyén outbreak’s relevance too broadly, given that passengers have dispersed globally. The forecast does not fully leverage the long resolution timeframe, which allows for both comprehensive asymptomatic case reporting and potential H2H cases to emerge and be officially linked.

Objections:

• The forecast’s initial assumption of 8 cases from WHO is inconsistent with the verified grounding fact of “at least 12” suspected or confirmed cases as of May 7-8, leading to an underestimated starting point for the distribution.
• The reasoning dismisses the potential for secondary human-to-human transmission (H2H) by overemphasizing the ‘contained-ship scenario.’ The grounding facts explicitly mention global dispersal of passengers and historical H2H transmission for Andes virus to household and healthcare contacts, which could lead to additional cases outside the initial onboard exposure.
• The probability distribution, particularly the mode at 13-14 and the 95th percentile at 25, seems too conservative considering the verified current count of at least 12 cases and the broad criteria for counting asymptomatic, explicitly reported cases from a pool of 147 passengers/crew.

Reviewer (fact-checker)

3 unsupported claim(s). Verdict: minor_issues

Unsupported claims:

• Historical Andes virus H2H clusters typically involve 3–16 cases
• if ~5–10% of 147 passengers/crew seroconvert and are officially reported, the total could rise into the 15–25 range
• Mode at 13–14, 80th-percentile ~19, 95th-percentile ~25

Adjustment Rationale:

Adjusted in two directions to reconcile reviewer concerns. (1) Modestly thinned the right tail (>25) because the reasoning reviewer correctly noted that systematic publication of asymptomatic serosurvey results within official cluster counts is uncommon — without a published serosurvey the path to 25+ requires substantial unobserved secondary transmission, which contradicts ECDC/WHO ‘low risk’ assessments and the closing incubation window. (2) Slightly raised mass in the 11–18 range because the calibration reviewer correctly noted there is a queue of pending cases (KLM attendant, three May 6 evacuees, French and Swiss contacts) that are likely to convert into qualifying classifications (confirmed/probable/suspected) over the ~3 months until August 1. Held firm on the floor at 8 rather than 12: the resolution criteria explicitly excludes PUIs, pending cases, and contacts under surveillance, so media-reported ‘12+’ includes individuals who do not yet meet qualifying categories. Held firm on overall shape: the central tendency (mode 13–14) is consistent with WHO’s current 8 plus likely classification of 4–6 pending cases. The right tail beyond 30 remains nonzero but small, reflecting genuine but unlikely scenarios (super-spreader-like dynamics or mass serosurvey publication).

forecast_sdk | 571.9s | research $1.1626 + reviewers $0.0127 + reconcile $0.1827 = $1.3580

Extracted Percentile Values: {0.1: 5.1, 1.0: 5.4, 5.0: 6.2, 10.0: 7.0, 15.0: 7.6, 20.0: 8.0, 25.0: 8.5, 30.0: 9.0, 35.0: 9.6, 40.0: 10.2, 45.0: 10.8, 46.0: 10.9, 47.0: 11.0, 48.0: 11.2, 49.0: 11.3, 50.0: 11.5, 51.0: 11.6, 52.0: 11.8, 53.0: 12.0, 54.0: 12.2, 55.0: 12.4, 60.0: 13.3, 65.0: 14.3, 70.0: 15.5, 75.0: 17.0, 80.0: 19.5, 85.0: 23.0, 90.0: 29.0, 95.0: 40.0, 99.0: 80.0, 99.9: 130.0}

Model’s Answer: The best anchor is the WHO update of May 7, which is more authoritative and more recent than the initial May 4 DON: 8 total publicly reported linked cases, of which 5 are confirmed and 3 suspected, with 3 deaths. That is the starting stock.

Step one is to separate what can still change. There are really three moving parts: some currently suspected cases could be ruled out, additional already-exposed passengers or crew could be found through contact tracing or testing, and limited secondary transmission from Andes virus could add a small number of downstream household or close-contact cases. Because the question counts suspected, probable, confirmed, and explicitly linked asymptomatic infections, improved testing can raise the total even if transmission is not extensive.

Step two is to weigh the time window. From May 8 to August 1 is long enough for essentially all primary onboard infections to declare themselves and for one more generation of close-contact transmission to be detected and publicly reported. It is not long enough for a large multi-generation epidemic to be the default, especially given WHO’s repeated “low risk” assessment and the known limited human-to-human transmission profile of Andes virus. So this looks much more like a bounded cluster than an exponential process.

Step three is to think about reporting behavior, because the metric is public reporting, not true infections. WHO and multiple national authorities are engaged, kits were distributed to five countries, and at least 12 countries were notified. That supports some upward revision from 8. But public authorities often do not publish every mild or asymptomatic linked infection in detail, especially once the acute media moment passes. That keeps me from pushing the median too high.

My base case is that the final publicly reported total lands modestly above the current 8, most likely in the low teens. I see the modal pathways as either a small number of added onboard or immediate-contact cases, or a slightly broader serologic/accounting exercise that lifts the total into the teens without a major transmission chain.

A low outcome would happen if most of the currently suspected cases are later ruled out and few or no additional linked infections are publicly announced. I put that low-end story at roughly a one-in-five chance. The floor is constrained by the already reported 5 confirmed, and practically the cluster is unlikely to end much below the high single digits unless multiple suspected cases are reversed.

A high outcome would happen if authorities do broad serology among passengers and crew, uncover several asymptomatic linked infections, or detect more than one secondary chain among close contacts after disembarkation. I put the genuinely high-end story, say above 30, at roughly a one-in-ten chance, and very large totals above 100 as extremely unlikely but not impossible enough to ignore under Brier-optimal forecasting.

The assumptions most likely to matter are these. First, that WHO’s “limited human-to-human transmission” framing remains descriptively accurate here; I assign about a 15% chance this underestimates secondary spread enough to materially raise the count. Second, that authorities will not conduct and publicly release a near-complete serologic census of passengers and crew; I assign about a 25% chance this assumption fails, which is the main reason for a fatter right tail. Third, that the current 8-case picture is not materially overstated by suspected cases later excluded; I assign about a 20% chance this assumption fails, which creates the limited left tail toward 5 to 7.

Overall, I center the distribution around 11 to 12, with a 90% interval roughly from 6 to 40 and a long but thin right tail.

Percentile .1: 5.1 Percentile 1: 5.4 Percentile 5: 6.2 Percentile 10: 7.0 Percentile 15: 7.6 Percentile 20: 8.0 Percentile 25: 8.5 Percentile 30: 9.0 Percentile 35: 9.6 Percentile 40: 10.2 Percentile 45: 10.8 Percentile 46: 10.9 Percentile 47: 11.0 Percentile 48: 11.2 Percentile 49: 11.3 Percentile 50: 11.5 Percentile 51: 11.6 Percentile 52: 11.8 Percentile 53: 12.0 Percentile 54: 12.2 Percentile 55: 12.4 Percentile 60: 13.3 Percentile 65: 14.3 Percentile 70: 15.5 Percentile 75: 17.0 Percentile 80: 19.5 Percentile 85: 23.0 Percentile 90: 29.0 Percentile 95: 40.0 Percentile 99: 80.0 Percentile 99.9: 130.0

Forecast: Median = 12.78 cases

• The current confirmed and suspected case count stands at 8, with several pending tests that could add 0-3 cases in the near term.
• Strong negative signals from contact tracing suggest that many of the pending tests are likely to return negative results, limiting potential case increases.
• The possibility of serological surveys identifying asymptomatic cases presents the largest uncertainty; if conducted and reported, this could significantly inflate the total case count.
• The biological window for new cases is closing soon, as most infections should manifest before late May, making it unlikely for many new cases to be reported after that.
• Reporting practices by various national authorities may influence the final count, especially if they choose to publish independent tallies before the August deadline.

Forecast (percentiles, aggregated):

• p10: 10.0
• p25: 13.0
• p50: 17.0
• p75: 26.0
• p90: 39.0

Per-model forecasts:

• 1: p10=9, p25=12, p50=17, p75=26, p90=42
• 2: p10=12, p25=15, p50=20, p75=30, p90=48
• 3: p10=9, p25=11, p50=15, p75=22, p90=32

Rationale: Summary The baseline of publicly reported cases linked to the MV Hondius cluster sits between 8 and 13 individuals as of early May 2026, comprising a mix of confirmed and suspected cases. The outbreak involves the Andes virus, a hantavirus distinct for its limited human-to-human transmission capabilities that generally requires close and prolonged contact. The primary exposed population is strictly bounded to the approximately 150 passengers and crew who were aboard the ship. Because the virus has an incubation period of up to eight weeks, additional cases may continue to emerge through June and early July as passengers disperse to various countries. A massive, uncontrolled spread is highly unlikely given the transmission dynamics, but a moderate increase in the case count is expected. The distribution of 2,500 diagnostic kits and intense international contact tracing will likely uncover additional mild, asymptomatic, or secondary household infections. The median estimate of 17 reflects a realistic but contained expansion from the current figures, accounting for these additional discoveries. The upper tail (P90 of 39) is constrained by historical precedents, roughly matching the 34 cases seen during the 2018-2019 Epuyén outbreak, which remains the largest documented person-to-person Andes virus event.

Strongest Arguments for Higher Values

• Broad inclusion criteria: Asymptomatic seropositive individuals count toward the final resolution if explicitly linked to the cluster. Systemic testing of dispersed passengers is highly likely to identify these silent infections.
• Lengthy incubation period: With an incubation window of up to eight weeks, secondary spread among close family members or contacts of disembarked passengers remains a live threat through mid-summer.
• Favorable aggregation rules: The resolution allows for the inclusion of unique cases reported by regional and national authorities as long as the WHO has not explicitly rejected them, potentially elevating the cumulative deduplicated count.

Strongest Arguments for Lower Values

• Transmission barriers: Andes virus does not spread easily through casual contact, making an explosive community outbreak beyond the ship’s passengers and their immediate caregivers highly improbable.
• Bounded primary exposure: The initial exposure event is entirely limited to a known, contained cohort of roughly 147 individuals.
• Reclassification of cases: Several currently “suspected” cases may test negative and be officially ruled out before August 1, which would actively reduce the current working count and drag down the final total.

Key Uncertainties

• The final epidemiological classification of the current suspected cases, heavily dependent on how many are definitively ruled out versus confirmed.
• The true prevalence of asymptomatic Andes virus infections among the exposed cruise-ship cohort, which could significantly inflate the numbers if widely tested.
• The degree to which disembarked passengers adhered to isolation protocols, which dictates the severity of localized secondary household clusters across the 12+ monitoring countries.