Sixty-Two Illnesses After a Wedding Buffet: Reading the Attack Rates
A wedding reception catered from a single licensed kitchen serves 180 guests a buffet of roasted chicken in cream sauce, potato salad, tossed greens, rolls and cheesecake. The chicken was roasted the previous afternoon, cooled in 6-inch pans on a speed rack at room temperature for about 3 hours, refrigerated, sliced at 8 a.m., and held in chafing dishes at 120–130°F until service at 1 p.m. Within 4 to 14 hours (median 10) of eating, 62 guests report watery diarrhea and severe abdominal cramps, with little vomiting and no fever; symptoms last about 24 hours. Interviews show 55 of 90 guests who ate the chicken became ill, versus 7 of 90 who did not. For potato salad, 40 of 110 eaters became ill versus 22 of 70 non-eaters. Three kitchen staff also report similar symptoms.
1. Calculate the attack rates and relative risks, and state which food the data implicate - including any reason for caution.
Read the worked answer
Chicken
- Attack rate, exposed: 55/90 = 61%
- Attack rate, unexposed: 7/90 = 7.8%
- Relative risk: 61.1 / 7.8 ≈ 7.9
- Attributable fraction in the exposed: (61.1 − 7.8)/61.1 ≈ 87%
Potato salad
- Exposed: 40/110 = 36%; unexposed: 22/70 = 31%
- RR ≈ 1.16 - essentially no association.
Interpretation: chicken is the strong vehicle; potato salad is not.
Reasons for caution:
- Confounding by co-consumption. Nearly everyone who ate potato salad probably also ate chicken, so potato salad's crude attack rate is dragged upward by the chicken eaters inside that group. A stratified analysis (potato salad within chicken eaters, and vice versa) is needed before dismissing it entirely.
- The 7 illnesses among non-chicken-eaters are not zero. That can reflect misclassification of what people ate (cream sauce, cross-contamination on the line, a shared utensil), or a second vehicle. It should be explained, not ignored.
- Buffet self-service makes "did you eat it" a soft exposure variable; portion and serving-line contamination matter.
2. Using incubation period, symptom profile and the preparation history, what agent do you suspect, and what alternative would you have to rule out?
Read the worked answer
Most likely: Clostridium perfringens.
Converging lines of evidence:
- Incubation: 4–14 hours, median 10 - squarely in the C. perfringens range (typically 6–24 h, often 8–12 h).
- Symptoms: watery diarrhea and severe cramps, minimal vomiting, no fever, ~24-hour duration. That is the classic C. perfringens picture; vomiting-predominant illnesses point elsewhere.
- Vehicle and handling: cooked meat, large batch, ~3 hours of slow ambient cooling, then marginal hot-holding at 120–130°F (below the 135°F holding threshold). That is the textbook setup for spore germination and rapid vegetative growth.
Alternatives to rule out:
- Bacillus cereus (diarrheal type) - same 8–16 h, diarrhea-predominant profile; usually linked to rice, grains, starches, but cannot be excluded on clinical grounds alone. Culture of stool and food distinguishes them.
- Salmonella - usually 12–72 h with fever and vomiting; the short median and afebrile course argue against it, but the chicken is a plausible vehicle and it must be tested for.
- Staphylococcus aureus / B. cereus (emetic) - 1–6 h, vomiting-dominant; inconsistent with this cluster.
- Norovirus - 12–48 h, vomiting prominent, secondary person-to-person spread; not the leading fit here.
The staff angle matters. Three ill food handlers with similar symptoms could mean a handler source, especially if their illnesses began before the event. That would broaden the investigation to bare-hand contact with ready-to-eat items and shift the working hypothesis. Interview them for onset times and duties first.
3. What immediate control and investigative actions do you take, and which are easy to get wrong?
Read the worked answer
Control now:
- Exclude or restrict ill food handlers per the employee health policy; obtain onset dates and job duties before excluding, because timing determines whether they are a source or a victim.
- Hold and embargo remaining product, including leftover chicken, sauce and any buffet items handled with shared utensils. Do not let the kitchen "clean up" by discarding evidence - photograph, label, and retain samples.
- Stop the practice, not just the batch: correct the cooling method (shallow pans, ice bath, blast chiller, documented time-temperature) and enforce 135°F minimum hot-holding. These are the operative failures.
Investigate:
- Collect food samples (chicken, sauce, potato salad) and clinical specimens. For C. perfringens, ask the laboratory specifically for anaerobic culture and spore enumeration - routine stool culture will miss it.
- Build a full 2×2 table for every menu item and stratify for co-consumption.
- Interview the kitchen about the cooling log: was the 3-hour ambient cooling documented, or is it reconstruction?
- Notify the public health authority; clusters are reportable even where individual cases are not.
Easy to get wrong:
- Discarding leftovers before sampling destroys the only direct evidence.
- Treating the buffet as a single exposure - self-service means each item needs its own attack rate.
- Assuming the kitchen staff are the cause when they are more likely additional cases from the same meal.