Coursework – Epidemiological Investigation of a Foodborne Outbreak

Introduction

Foodborne illness imposes a substantial and largely preventable burden on the Australian health system, with an estimated 4.1 million cases of domestically acquired foodborne gastroenteritis each year (Kirk et al., 2014). Salmonellosis accounts for a disproportionate share of the resulting hospitalisations, and national surveillance has documented a sustained rise in notification rates since 2000, with Queensland consistently above the national average (Ford et al., 2016; OzFoodNet Working Group, 2023). Catered functions recur as settings for point source outbreaks because a single contaminated item reaches a large, precisely enumerable group within a narrow window of time.

This coursework applies the ten step outbreak investigation framework to a hypothetical outbreak of salmonellosis following a catered fundraising dinner in the Darling Downs region of southern Queensland in June 2025. It constructs a case definition, describes the outbreak by person, place and time, calculates food specific attack rates and relative risks to identify the vehicle, specifies control measures and notification pathways, and appraises the limitations of the investigation. All figures are invented but internally consistent.

Outbreak Context and Initial Notification

On Monday 16 June 2025 a general practitioner contacted the regional public health unit to report four patients presenting since the previous evening with fever, cramping abdominal pain and profuse diarrhoea, three of them with blood in the stool. All four had attended a catered fundraising dinner at the Glenview Function Centre on the evening of Saturday 14 June, at which 148 guests were served. Two faecal specimens had already yielded Salmonella enterica serovar Typhimurium, triggering the pathology laboratory’s independent notification obligation under the Public Health Act 2005 (Qld). A team of clinical, nursing, epidemiological and local government environmental health staff was convened that afternoon.

Applying the Ten Step Framework

Outbreak investigation is conventionally described as a sequence of ten steps, although in practice several proceed concurrently and control measures are never deferred until the analysis is complete (Gregg, 2008; World Health Organization [WHO], 2008). Table 1 summarises the framework and its application here; the sections that follow develop the analytically substantial steps.

Table 1: The ten steps of outbreak investigation applied to the Glenview Function Centre outbreak

Step 2 warrants quantification. The district records a mean of 4.2 salmonellosis notifications per week during June across the preceding five years. Thirty one were received in the week commencing 15 June 2025, giving an observed to expected ratio of 31 / 4.2 = 7.4. A sevenfold excess establishes an outbreak rather than chance fluctuation in endemic disease (Webb et al., 2020).

Case Definition

A case definition standardises ascertainment across investigators and embeds a trade off between sensitivity and specificity (Gregg, 2008). A broad definition was applied during active case finding so that cases were not missed, then narrowed for the analytical phase so that the cohort was not diluted by unrelated gastroenteritis. The definitions adopted are structured by person, place and time.

• Confirmed case: a person who attended the function at the Glenview Function Centre on 14 June 2025 and from whom Salmonella Typhimurium was isolated from a faecal specimen collected between 15 and 30 June 2025.
• Probable case: a person who attended the function and experienced three or more loose stools within any 24 hour period, or fever with abdominal cramping, with onset 6 to 96 hours after the meal, without laboratory confirmation.
• Exclusions: persons whose symptoms began before the meal was served, and persons who did not attend. Symptomatic household contacts were investigated separately as possible secondary cases.

The 96 hour bound reflects the accepted incubation period for non typhoidal salmonellosis of 6 to 72 hours, extended to capture late onsets (Communicable Diseases Network Australia [CDNA], 2022). Applied to the 132 attendees interviewed, these definitions identified 47 cases, 19 confirmed and 28 probable.

Descriptive Epidemiology

Time

Figure 1 shows the classic signature of a point source exposure: a steep rise, a single peak on the first full day after the function, and a rapid decline with no secondary wave among attendees. The shortest incubation period was 9 hours and the longest 77 hours, with a median of 26 hours. That distribution fits non typhoidal Salmonella and excludes a preformed toxin such as that of Staphylococcus aureus, where onset is measured in hours rather than days (CDNA, 2022).

Place

All 47 cases had attended the single function, and none arose among diners in the venue’s public bistro that evening, which used a separate menu and preparation bench. Cases were spread across 14 of the 15 guest tables, and a chi square test of case distribution by seating was not significant, arguing against both a table specific serving error and person to person transmission at the venue. Forty one cases resided in the shire and six had travelled from Toowoomba or Brisbane, which carried implications for interjurisdictional notification.

Person

Cases ranged in age from 4 to 81 years, with a median of 44 years. Twenty six cases were female and 21 male, giving attack rates of 26 / 71 = 36.6 per cent among female attendees and 21 / 61 = 34.4 per cent among male attendees, a relative risk of 1.06 and no meaningful difference. Diarrhoea was universal (47 of 47), followed by abdominal cramps (43, 91.5 per cent), fever (38, 80.9 per cent), nausea (31, 66.0 per cent), vomiting (18, 38.3 per cent) and blood in the stool (11, 23.4 per cent). Median symptom duration was five days, six cases were hospitalised (6 / 47 = 12.8 per cent), all at the extremes of age, and there were no deaths. Three household contacts who had not attended fell ill between 20 and 22 June and were classified as secondary cases.

Hypothesis Generation and Analytical Epidemiology

A median incubation of 26 hours, invasive symptoms and a confirmed Salmonella isolate narrowed the hypothesis space considerably. Menu review identified one item made from raw shell eggs, an aioli served with the seafood entree. Raw egg products dominate identified vehicles in Australian S. Typhimurium outbreaks (Moffatt et al., 2016), so the aioli became the primary hypothesis while five other items were retained for comparison.

Because the population at risk was fully enumerable from the guest list, a retrospective cohort study was the appropriate design; a case control study becomes necessary only when that population cannot be defined (Webb et al., 2020). Interviews were completed with 132 of 148 attendees, a response fraction of 132 / 148 = 89.2 per cent. Two measures were calculated for each item:

• Attack rate (AR) = (number ill among the exposed / total exposed) × 100
• Relative risk (RR) = AR among the exposed / AR among the unexposed

The overall attack rate was 47 / 132 × 100 = 35.6 per cent. Table 2 presents the food specific results.

Table 2: Food specific attack rates and relative risks for six menu items, Glenview Function Centre, 14 June 2025 (n = 132)

Note. Every row sums to 132 respondents and 47 cases. AR = attack rate; RR = relative risk; CI = confidence interval.

The calculations for the implicated item are set out step by step.

• AR among the exposed = 44 / 78 × 100 = 0.564 × 100 = 56.4 per cent
• AR among the unexposed = 3 / 54 × 100 = 0.056 × 100 = 5.6 per cent
• RR = 56.4 / 5.6 = 10.2
• Risk difference = 56.4 – 5.6 = 50.8 percentage points

The same procedure applied to the prawn cocktail yields AR exposed = 41 / 96 × 100 = 42.7 per cent, AR unexposed = 6 / 36 × 100 = 16.7 per cent and RR = 42.7 / 16.7 = 2.6. Two derived measures aid interpretation. The attributable fraction among the exposed is (RR – 1) / RR = (10.2 – 1) / 10.2 = 0.902, so about 0.902 × 44 = 40 of the 44 cases among aioli consumers are attributable to that exposure. The population attributable fraction is (AR total – AR unexposed) / AR total = (35.6 – 5.6) / 35.6 = 0.843, so removing the aioli would have prevented roughly 84 per cent of illness in the cohort. No other item produced a significant excess, and the roast beef relative risk below unity reflects that abstainers from the main course had still eaten the entree.

Controlling for Co-consumption

The elevated relative risk for the prawn cocktail required scrutiny, because the aioli was served with that entree and the two exposures are heavily correlated. Stratification is the standard means of separating a genuine effect from confounding by a co-consumed item (Rothman et al., 2021), and Table 3 sets out the result.

Table 3: Attack rate for the prawn cocktail entree stratified by raw egg aioli consumption

Within both strata the association between the prawn cocktail and illness collapses towards the null (RR = 1.1 and RR = 0.5) against a crude estimate of 2.6, so the entree was a confounded marker of aioli consumption rather than a vehicle. Reversing the stratification confirms the reading: among prawn cocktail eaters the relative risk for aioli is 57.1 / 3.8 = 15.0, and among those who declined the entree it is 50.0 / 7.1 = 7.0. The aioli association persists strongly within strata, which is the pattern expected of a genuine causal exposure.

Laboratory and Environmental Findings

Statistical association alone is not confirmation, so step 9 required reconciliation with microbiological and environmental evidence. All 19 clinical isolates shared a single multiple locus variable number tandem repeat analysis profile, and whole genome sequencing placed them within three single nucleotide polymorphisms of one another, consistent with a common source. The same profile was recovered from a refrigerated leftover portion of aioli and from pooled unwashed shell eggs collected at the supplying producer. Environmental health officers established that the aioli had been made on the morning of 14 June from roughly 30 raw shell eggs and held on a bench at about 14 degrees Celsius for five hours before service, contrary to Standard 3.2.2 of the Australia New Zealand Food Standards Code, which requires potentially hazardous food to be kept at or below 5 degrees Celsius (Food Standards Australia New Zealand [FSANZ], 2021). This convergence of epidemiological, laboratory and environmental evidence meets the usual threshold for identifying a vehicle (WHO, 2008).

Control Measures and Notification

Control measures were implemented progressively from 16 June rather than at the conclusion of the analysis. Immediate actions comprised withdrawal of all remaining aioli, an embargo on the implicated egg consignment, sanitation of preparation surfaces, and exclusion of two symptomatic food handlers under the criteria for high risk settings, which require freedom from symptoms for at least 48 hours before a return to food handling (CDNA, 2022). Attendees received advice on hydration, hand hygiene and avoiding food preparation for others while symptomatic.

Regulatory action proceeded under the Food Act 2006 (Qld), administered locally by council environmental health officers, who issued an improvement notice requiring the venue to cease using raw shell eggs in ready to eat foods, substitute a pasteurised egg product and revise its food safety program. The producer was referred to Safe Food Production Queensland, the accrediting body for the state egg food safety scheme, for audit of on farm hygiene and traceability (Safe Food Production Queensland, 2023).

Salmonellosis is notifiable by both pathology laboratories and medical practitioners under the Public Health Act 2005 (Qld). Notifications reach the regional public health unit, which manages the response in line with state outbreak guidelines (Queensland Health, 2022) and forwards data to the Communicable Diseases Branch. The outbreak was reported to OzFoodNet, which collates outbreak summaries for the National Notifiable Diseases Surveillance System (OzFoodNet Working Group, 2023) and, because six attendees lived outside the district, provided the mechanism for alerting other jurisdictions to the subtype.

Limitations

Several limitations qualify these findings. Interviews occurred six days after the meal, so recall bias is plausible and may be differential, since unwell respondents typically reconstruct a meal more thoroughly than well respondents, which tends to inflate relative risks. Sixteen of 148 attendees (10.8 per cent) were not interviewed, and if non-respondents were less likely to have been ill, the overall attack rate of 35.6 per cent is overstated. Precision is uneven: only 11 respondents avoided the bread rolls, so that estimate carries a very wide interval. Testing six exposures inflates the family wise error rate, although a relative risk of 10.2 with a lower bound of 3.3, supported by biological plausibility and matching isolates, is not credibly a multiple comparisons artefact.

Stratification addressed confounding by the co-consumed entree, but residual confounding cannot be excluded, and logistic regression would be preferable where many correlated items appear on a menu. Consumption was recorded dichotomously, so no dose response relationship could be tested, a meaningful omission given that portion size predicts infection risk. Environmental sampling occurred four days after preparation, so the leftover aioli may not represent the batch served. The investigation proceeded under public health legislation rather than as research, so formal ethics review was not required, although confidentiality obligations applied throughout.

Conclusion

Systematic application of the ten step framework identified raw egg aioli as the vehicle of a point source outbreak of S. Typhimurium affecting 47 of 132 interviewed attendees at a catered function in regional Queensland. The epidemic curve, incubation distribution and symptom profile were consistent with non typhoidal salmonellosis, and the cohort analysis produced a relative risk of 10.2 for the aioli, a population attributable fraction of 84.3 per cent, and no competing association once co-consumption was controlled by stratification. Laboratory subtyping and environmental inspection converged on the same conclusion. The episode illustrates a persistent and preventable Australian pattern in which raw shell eggs are used in ready to eat foods and then held outside temperature control (Moffatt et al., 2016). The durable preventive measure is therefore not better refrigeration but substitution of pasteurised egg product in any dish that will not be cooked, reinforced by the traceability arrangements already in place for Queensland egg producers.

References

Communicable Diseases Network Australia. (2022). Salmonellosis: CDNA national guidelines for public health units. Australian Government Department of Health and Aged Care.

Food Act 2006 (Qld).

Food Standards Australia New Zealand. (2021). Safe food Australia: A guide to the Food Standards Code (4th ed.).

Ford, L., Glass, K., Veitch, M., Wardell, R., Polkinghorne, B., Dobbins, T., Lal, A., & Kirk, M. D. (2016). Increasing incidence of Salmonella in Australia, 2000-2013. PLOS ONE, 11(10).

Gregg, M. B. (Ed.). (2008). Field epidemiology (3rd ed.). Oxford University Press.

Kirk, M. D., Ford, L., Glass, K., & Hall, G. (2014). Foodborne illness, Australia, circa 2000 and circa 2010. Emerging Infectious Diseases, 20(11), 1857-1864.

Moffatt, C. R. M., Musto, J., Pingault, N., Miller, M., Stafford, R., Gregory, J., Polkinghorne, B. G., & Kirk, M. D. (2016). Salmonella Typhimurium and outbreaks of egg-associated disease in Australia, 2001 to 2011. Foodborne Pathogens and Disease, 13(7), 379-385.

OzFoodNet Working Group. (2023). Monitoring the incidence and causes of diseases potentially transmitted by food in Australia: Annual report. Communicable Diseases Intelligence, 47.

Public Health Act 2005 (Qld).

Queensland Health. (2022). Guidelines for the investigation of gastroenteritis and suspected foodborne disease outbreaks. Queensland Government.

Rothman, K. J., Lash, T. L., & VanderWeele, T. J. (2021). Modern epidemiology (4th ed.). Wolters Kluwer.

Safe Food Production Queensland. (2023). Egg food safety scheme: Annual report. Queensland Government.

Webb, P., Bain, C., & Page, A. (2020). Essential epidemiology: An introduction for students and health professionals (4th ed.). Cambridge University Press.

World Health Organization. (2008). Foodborne disease outbreaks: Guidelines for investigation and control.