Industries / Retail, hospitality and personal services / Hospitality and tourism / Cruise lines
Health and outbreaks
This work keeps infectious disease from spreading among thousands of people living close together on a ship, and treats those who fall ill far from a hospital.
7
Milestones achieved, from 1975 to 2026.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1975
Ship inspections begin
After many outbreaks of stomach illness on cruise ships in the early 1970s, the US Centers for Disease Control and Prevention (CDC) set up the Vessel Sanitation Program (VSP) in 1975. It covers ships that carry 13 or more passengers, sail a foreign itinerary and call at US ports. Each gets two unannounced inspections a year, scored out of 100; a score below 86 fails and brings a second unannounced inspection. A 1989 review found that no ship passed when the program began, and that on cruises of 3 to 15 days outbreaks fell from 8.1 to 3.0 per 10 million passenger days between 1975–79 and 1980–85.
Credit
CDC Vessel Sanitation Program
1995
Guidelines for ship clinics
No government agency regulates medical practice on cruise ships, the CDC says. In 1995 the American College of Emergency Physicians (ACEP) published consensus guidelines for cruise ship medical facilities, and it updated them in 2023. By the CDC's account about 95 percent of illnesses and injuries seen on board are treated on the ship, and the rest may need evacuation to care on shore.
Credit
American College of Emergency Physicians
2000
Illness counted before arrival
CDC started the Maritime Illness Database and Reporting System in 2000. Ships report their running count of stomach illness 24 to 36 hours before reaching a US port and update it 4 hours out. CDC posts an outbreak publicly when 3 percent or more of passengers or crew report symptoms. Across 37,258 voyage reports from 252 ships between 2006 and 2019, passenger illness fell from 32.5 to 16.9 cases per 100,000 travel days, and about 90 percent of outbreaks with a known cause were norovirus.
Credit
CDC Vessel Sanitation Program (Keisha A. Jenkins and colleagues, 2021 report)
2018
Scores fail to predict outbreaks
Christopher James Taylor compared the 1,197 VSP inspections published from 2012 to 2017 with the 50 stomach illness outbreaks in the same years. Inspections followed by an outbreak had scored no lower than the rest: an average of 96.4 against 95.1 out of 100, a difference small enough to be chance.
LessonA high sanitation score did not mean a lower risk of an outbreak; Taylor concluded the scores have "no prognostic value".
Credit
Christopher James Taylor, in International Maritime Health
2020
Quarantine aboard Diamond Princess
The Diamond Princess was quarantined in Yokohama, Japan, on 3 February 2020. Passengers were confined to their cabins from 5 February, while crew kept working. Of 3,711 people on board, 712 (19.2 percent) tested positive for the virus that causes COVID-19, 331 of them without symptoms when tested, and nine of them died. Most passenger infections happened before the quarantine, but crew infections peaked after it began. Traces of the virus's genetic material were found on cabin surfaces up to 17 days after people left.
LessonConfining passengers to cabins did not protect the crew, who could not isolate while they kept the ship running.
Credit
CDC investigators Leah F. Moriarty and colleagues
2020
US cruising halted
On 14 March 2020 the CDC issued a No Sail Order for cruise ships. On 30 October 2020 a Conditional Sailing Order replaced it with a phased path back to sailing, whose later phases were released in April and May 2021. Cruising from US ports restarted in summer 2021, after changes to the order, a lawsuit by the state of Florida and the arrival of vaccines, by Cruise Industry News' account. The order expired on 15 January 2022, and the voluntary program that followed ended on 16 July 2022.
Credit
CDC
2026
Hantavirus on an expedition ship
The Dutch-flagged expedition ship MV Hondius left Ushuaia, Argentina, on 1 April 2026. A passenger who had travelled for more than three months in South America fell ill on 6 April and died on board on 11 April; the World Health Organization (WHO) was told of the cluster on 2 May. WHO said the first case was most probably infected on land in Argentina and that the evidence pointed to person-to-person spread on board. By 9 July, 13 cases of Andes hantavirus (12 confirmed) had been linked to the ship, 3 of them fatal, among 10 of 121 passengers and 3 of 61 crew, and the response covered people of 23 nationalities.
LessonReviewers found the existing rules left unsettled which ports a remote ship with a dangerous infection could use, and how to evacuate and trace its passengers.
Credit
WHO; the international investigation team writing in Eurosurveillance
Watch for
Sewage testing at the berth
Researchers at the University of Pisa tested raw and treated sewage from 10 cruise ships at an Italian port between July and October 2024 for 10 viruses. Norovirus genogroup II turned up in every raw sample and the COVID-19 virus in 60 percent. The COVID-19 virus was found in the sewage on about 85 percent of arrivals whose ship had diagnosed cases. Some ships' treatment plants cut virus levels by less than a hundredfold, and their outflow carried more bacteria than the benchmark in International Maritime Organization (IMO) guidelines.
Achieved whena port health authority tests ship sewage routinely for a full season and publishes the results, or another group repeats the match with more ships.
Claimed by
Ileana Federigi and colleagues, University of Pisa, April 2026
Open
01
Which faults come before outbreaks
A dataset of every itemized fault in VSP inspection reports since 2012, each sorted into a fixed category (hand-washing stations, food temperatures, ice machines, pools and hot tubs, and so on) and linked to the outbreaks CDC posted afterwards, with a test of whether any category, rather than the total score, comes before outbreaks more often than chance.
The briefCredit · unclaimed02
Messages that speed sick reports
A randomized online experiment testing which messages make cruise passengers say they would report stomach symptoms to the ship's medical centre sooner. Candidates include a free consultation for stomach illness, a plain account of how one sick person infects others, and a reassurance about what isolation involves. Medical care on board usually costs extra, the CDC notes.
The briefCredit · unclaimed
Open 01 · the brief
Which faults come before outbreaks
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A dataset of every itemized fault in VSP inspection reports since 2012, each sorted into a fixed category (hand-washing stations, food temperatures, ice machines, pools and hot tubs, and so on) and linked to the outbreaks CDC posted afterwards, with a test of whether any category, rather than the total score, comes before outbreaks more often than chance.
Why it is still open
Taylor's 2018 study tested only total scores and found they did not predict outbreaks. Mark Silverberg collected the itemized faults from CDC's inspection site into a public dataset in February 2016, and the repository has not changed since. A 2026 analysis of 425 CDC outbreak investigations on GitHub uses the outbreak records only, not the inspections.
Built with what exists
CDC's public Inspection Query Tool; Python to collect the reports; a language model to sort the free-text fault notes into categories, such as Claude (see the Note "Seven more flagged") or Google's Gemini Decision 1, which scores fixed-answer questions in bulk (see the Note "A thousand questions, one pass"); a statistics package such as R.
First step
Collect one year of reports with their fault lists, and hand-code 50 faults to settle the categories.
Done looks like
all published reports from 2012 to 2026 collected; the model's categories checked against hand coding on 200 randomly chosen faults with at least 95 percent agreement; the analysis plan registered in advance, and the dataset and results published.
Open 02 · the brief
Messages that speed sick reports
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A randomized online experiment testing which messages make cruise passengers say they would report stomach symptoms to the ship's medical centre sooner. Candidates include a free consultation for stomach illness, a plain account of how one sick person infects others, and a reassurance about what isolation involves. Medical care on board usually costs extra, the CDC notes.
Why it is still open
In a model of one Mediterranean norovirus outbreak, the 10 percent of infected people who caused 57 percent of onward cases waited longer before diagnosis (83 hours against 47), and shorter delays cut cases further (Alessandro De Bellis and colleagues, Journal of Travel Medicine, 2025). A 2024 review of 45 norovirus outbreaks on cruise ships found that only three of 13 studies had looked at travellers' behaviour, and called for interventions to improve symptom reporting.
Built with what exists
a survey panel such as Prolific; advance registration on the Open Science Framework; scenario questions checked by a ship doctor; translation into the main passenger languages.
First step
Write the scenario and the messages, and register the study on the Open Science Framework.
Done looks like
at least 2,000 adults who have taken a cruise, each randomly shown one of at least four messages or none, with the share who say they would report within 12 hours of symptoms compared across groups and published with the data.