Wedding Feast Diarrhoea Cluster in a Hill District
A hill district of 480,000 people reports a cluster of acute diarrhoea after a wedding feast on 14 June. Of 212 guests traced, 74 meet the case definition (three or more loose stools in 24 hours, or fever with abdominal cramps, with onset 4–48 hours after the meal). The median incubation is 18 hours. The kitchen was a temporary outdoor setup; water came from a stored tank, and the menu included chicken curry, paneer, rice and a raw vegetable salad. Interviews show 150 guests ate chicken curry (68 ill) and 62 did not (6 ill); 120 ate paneer (44 ill) and 92 did not (30 ill). Nine stool samples have been sent. Two food handlers report recent finger sepsis. The district has one microbiology laboratory 90 km away and a single ambulance. The wedding season continues for three weeks with several large events already booked with the same caterer.
1. Calculate the attack rates and risk ratios for the chicken curry and the paneer, and state what the findings do and do not establish.
Read the worked answer
Overall attack rate: 74/212 = 34.9%. Chicken curry: 68/150 = 45.3% among eaters versus 6/62 = 9.7% among non-eaters, giving a risk ratio of 4.7 (approximate 95% CI 2.1–10.3). Paneer: 44/120 = 36.7% versus 30/92 = 32.6%, risk ratio 1.1 - no meaningful difference. The chicken curry is the leading suspect; the paneer is not. The strength of the association and the tight incubation (median 18 hours) fit an invasive bacterial agent such as Salmonella more than a preformed toxin, but that is inference, not proof.
Limits: association is not causation. Guests who ate chicken may share another exposure, and menu items were served together, so stratified or multivariable analysis is needed. There is recall bias and misclassification of mild cases; portions were not quantified, so no dose-response can be assessed. The salad, the tank water or an infected handler could be the real vehicle, with chicken simply a marker; cross-contamination from raw chicken to salad would produce the same pattern. The 74 cases are only those traced, so the outbreak may be larger.
The defensible conclusion is that chicken curry is strongly associated and should focus environmental sampling and control, not that chicken has been proven to cause the outbreak.
2. Set out your investigation and control plan for the next 48 hours, given that the laboratory is 90 km away and further weddings are booked.
Read the worked answer
First 24 hours: active case finding through facility logs, pharmacy sales, absenteeism records and the guest list; build a line list with time, place, person, food items, portions and symptoms; update the epidemic curve. Collect stool from early cases in Cary-Blair and vomit if available; interview cooks and inspect the kitchen layout, water tank, hand-wash facilities, refrigeration and leftover food. Sample leftovers and tank water, and take handler stool and hand swabs. The distance to the laboratory is a logistics problem, not a reason to delay: send specimens cold in one trip and tell the laboratory what to look for.
Control before results: exclude symptomatic handlers until 48 hours symptom-free and stool-negative; suspend the caterer from new events until inspection is passed, because repeat exposures are booked. Advise boiling or chlorinating the tank water, provide hand-wash stations with soap, separate raw chicken preparation, thorough cooking and rapid cooling. Issue a public advisory with the case definition and where to obtain oral rehydration salts. Alert clinics to report bloody diarrhoea, dehydration or suspected invasive disease.
Also run a retrospective cohort analysis by guest group to test confounding, trace any earlier events served by the same caterer, and document all decisions. A precautionary suspension is justified when the same caterer is due to serve large gatherings and the epidemiological signal is strong, even before microbiology confirms it.
3. A journalist asks for the names of the two food handlers and a list of guests who fell ill. Draft your response and justify it.
Read the worked answer
Release the public-health facts: the number of cases linked to the meal, the food under investigation, the symptoms, the actions taken and where to seek care. Do not release names, addresses or other identifiers of cases or handlers.
Reasons: notification data are held under public health law for control, not for public exposure. The two handlers have finger sepsis but are not proven to be the source; naming them would defame them and could expose them to harassment or loss of livelihood. Identifying cases may deter others from reporting symptoms and seeking care, which damages the outbreak response. The public interest is served by warning about the caterer and the venue, not by individual identities. If a public warning is legally required, name the business only if the law permits and the risk justifies it.
Confirm the legal basis with the district legal officer, inform the handlers privately about exclusion, testing and their rights, and keep the line list under restricted access with a log of disclosure requests. If a statutory duty to disclose exists, release the minimum necessary and tell the affected people first where feasible. The ethical test is whether disclosure prevents harm that cannot be prevented otherwise; here it does not.