Potluck Gastroenteritis: Advisory Before Confirmation
On 14 June, a county health department receives reports of acute gastroenteritis among attendees of a 12 June potluck. The sign-in sheet lists 68 people; 41 meet a preliminary case definition of vomiting or at least 3 loose stools within 48 hours of the event. Median onset is 18 hours (range 6–36). The menu included roast chicken, potato salad, green salad, rice, and chocolate cake. The chicken was cooked 24 hours ahead, held in a garage at 18–22°C for 6 hours before serving, then left out during the 3-hour meal. Potato salad was kept on ice until serving. Two food handlers deny symptoms. Stool samples from three cases are pending; the laboratory has a 48-hour turnaround. The department must decide whether to issue a public advisory before confirmation and how to prioritize interviews during the first 24 hours. A local newspaper has already called.
1. Draft a case definition for this outbreak and explain how your choice of time window and clinical criteria affects sensitivity and specificity.
Read the worked answer
A workable definition is: a person who attended the 12 June potluck and developed vomiting or at least three loose stools in a 24-hour period, with onset 6–48 hours after the event. This combines person, place, time, and clinical criteria.
The 6–48-hour window is a deliberate compromise. A very short window (for example, 1–6 hours) would be specific for preformed toxins such as Staphylococcus aureus or Bacillus cereus, but it would miss slower pathogens such as Salmonella or Campylobacter. A very wide window (for example, 1–7 days) would raise sensitivity but pull in unrelated illness among 68 attendees, lowering specificity. Because the median onset is 18 hours, a 6–48-hour window captures the observed cluster while still allowing for several plausible agents.
The clinical criterion also trades off sensitivity and specificity. Requiring at least three loose stools in 24 hours is more specific than “any diarrhea” but may exclude mild cases; adding vomiting improves sensitivity for toxin-mediated illness. During the initial investigation, a slightly broader definition is appropriate to find cases; it can be narrowed later if the laboratory identifies a pathogen. The definition should be applied consistently to all 68 attendees and documented, because changing it mid-investigation complicates attack-rate comparisons.
2. Among 45 attendees who ate roast chicken, 34 became ill; among 23 who did not, 7 became ill. Calculate attack rates and the risk ratio, and state what the result does and does not establish.
Read the worked answer
Attack rate among chicken eaters: 34 ÷ 45 = 0.756, or 75.6%. Attack rate among non-eaters: 7 ÷ 23 = 0.304, or 30.4%. Risk ratio: 75.6% ÷ 30.4% = 2.49.
The result establishes a strong association between eating roast chicken and illness in this cohort. It does not, by itself, establish that the chicken was the vehicle. The non-eaters also had a 30% attack rate, which suggests either another exposure, secondary spread, or misclassification of food history. With only 68 people, the confidence interval around the risk ratio is wide, and recall bias may affect who remembers eating chicken. The chicken’s handling-cooked a day ahead, held at 18–22°C for 6 hours-is biologically plausible for toxin-producing bacteria, but laboratory confirmation or a more detailed food-specific analysis is still needed. The correct interpretation is “priority hypothesis,” not “confirmed cause.”
3. The incident commander asks whether to name roast chicken in a public advisory before stool cultures return. Give your recommendation with reasoning and one concrete wording.
Read the worked answer
Recommendation: issue an advisory now, but do not state that the chicken is confirmed as the cause. The cluster is large enough (41 ill), the exposure is recent, and people may still have leftover chicken at home. Waiting 48 hours for cultures delays actionable advice. At the same time, naming a specific food as confirmed before laboratory results can unfairly damage a vendor and may be wrong if another food or a food handler is responsible.
The advisory should be framed as precautionary and symptom-based: advise anyone with vomiting or diarrhea after the potluck to seek care and report illness; advise attendees not to eat leftover food from the event; and state that roast chicken is under investigation. A concrete wording is: “Health officials are investigating illness among potluck attendees. Until laboratory results are available, do not eat any leftover food from the event, especially the roast chicken, which is one focus of the investigation. Anyone with vomiting or diarrhea should contact the health department.” This gives the public a clear action without overstating certainty.