Applied epidemiology: Investigation of outbreaks of foodborne illness

Print Friendly, PDF & Email

Recent, high-profile foodborne illness outbreaks have created confusion and concern for Canadian consumers. From a Salmonella outbreak in western Canada linked to cucumbers in early fall­­, to the current E. coli outbreak linked to romaine lettuce, public health is working overtime to identify and solve illness outbreaks linked to the food supply. This post will provide an overview of the investigation of foodborne illness outbreaks in Canada.

First, it should be noted that foodborne (enteric) illness is a common occurrence, with an estimated four million episodes of domestically acquired cases occurring each year in Canada (Thomas, Murray, Flockhart, Pintar, Pollari, Fazil et al., 2013). Although often self-limiting, some of these cases require hospitalization, and a few lead to death. One estimate counts 11,600 hospitalizations and 238 deaths associated with domestically acquired foodborne illnesses in Canada each year (Thomas, Murray, Flockhart, Pintar,Fazil, Nesbitt et al, 2015). A variety of pathogens can cause enteric illness, including bacteria such as Salmonella, E.coli, campylobacter, shigella, and listeria; parasites such as giardia, cryptosporidium, and cyclospora, and viruses such as Hepatitis A and norovirus. Incubation periods for these pathogens range from 6 hours (salmonellosis minimum) to 70 days (listeriosis maximum) (Public Health Agency of Canada 2011, 2016).

Public health authorities do not investigate each of the 4 million cases of foodborne illness that occur in Canada each year. How then, do they identify outbreaks in the first place? There are 10 basic steps in an outbreak investigation, and the first step deals with this question. Let’s walk through the steps using a fictional, multi-provincial outbreak of E.coli in ground beef as an example (data is fabricated).

10 Steps of an Outbreak Investigation

1. Verify existence: Are the observed numbers of cases greater than expected? Consider seasonal variations, changes in reporting, new lab tests etc. Verotoxigenic (VTEC) E. coli is nationally notifiable in all provinces and territories. At the federal level, the National Enteric Surveillance Program (NESP) and the Canadian Notifiable Disease Surveillance System are two systems used to monitor cases of food-borne illness (Public Health Agency of Canada, 2017). NESP produces weekly reports comparing the reported incidence of illness to past averages, which can be useful to identify trends (Government of Canada, 2018).

In our fictional outbreak, for week 36 of 2018, NESP reported 16 cases of E. coli O157 nationally, but only expected to see 9. A closer look reveals 10 cases in Alberta, 5 in Saskatchewan, and 1 in Manitoba. We decided to investigate further.       

2. Confirm the diagnosis: What pathogen is causing illness? It is important to be as specific as possible.

For outbreak detection, it may not be enough to know the serotype and geography of a case of illness. The National Microbiology Laboratory (NML) provides reference services for strain identification and characterization, national laboratory-based surveillance,and dissemination of information through PulseNet Canada and NESP (Public Health Agency of Canada, 2018). The Centre for Foodborne, Environmental and Zoonotic Infectious Diseases (CFEZID) collaborates with NML to evaluate genetically linked multi-jurisdictional clusters of illness.

In our fictional outbreak, NML identified a cluster of 8 cases of E. coli O157 in Alberta and Saskatchewan. These cases were closely related as indicated by whole-genome sequencing (WGS). We decided to investigate further.  

3. Assemble team: Who do you need on your team to successfully investigate an outbreak?

Consider including local, provincial, and/or federal public health authorities, food safety partners, agriculture, laboratory, animal health specialists, communications, etc. CFEZID does regular follow-up on clusters of interest, and when a cluster meets certain criteria, the Outbreak Management Division (OMD) at the Public Health Agency of Canada (PHAC) initiates a nationally coordinated investigation (Government of Canada, 2018). Federal (PHAC, Health Canada, the Canadian Food Inspection Agency) and provincial/territorial partners come together and assess the cluster. If they feel it is worth investigating, the Foodborne Illness Outbreak Response Protocol is activated, and the partners form an Outbreak Investigation Coordinating Committee (OICC) (Government of Canada, 2018).

In our fictional outbreak, PHAC held an OICC assessment call with federal partners (CFIA, HC) and provincial public health authorities from Alberta and Saskatchewan 2 weeks after the first cases were reported in NESP. Since then, 10 additional cases were reported in NESP, including 6 analyzed by NML. Collaboratively, partners decided to activate the FIORP and form an OICC for collaborative investigation of the outbreak.

4. Define and find cases: Create a case definition and determine the extent of the outbreak.

The investigative team must agree on a case definition, which contains standard criteria for determining if individuals should be part of the outbreak or not. The case definition includes elements of person, place, time, and aetiology. Methods for case-finding in foodborne outbreaks can be public health related (ensuring complete reporting) and lab-based (checking for genetic matches in other jurisdictions).   

In our fictional outbreak, the investigative team drafted a case definition which stated that in order to be included in the current outbreak, cases of E. coli O157 had to be contracted within Canada, be genetically linked to the current outbreak by WGS, and have an illness onset of August 1, 2018 or later. For case finding, PHAC posted a Public Health Alert to the Canadian Network for Public Health Intelligence.   

5. Orient the data: Describe the cases in terms of person, place, and time.

Describe the distribution of cases by demographic indicators, including age, sex, and geography, as well as clinical indicators such as hospitalizations and complications. An important tool for characterizing outbreaks is the epidemiologic curve, which is a histogram of cases plotted by time of illness onset. The epi curve is a visual depiction of the timing and magnitude of the outbreak, as well as the pattern of transmission.  

In our fictional outbreak, the investigative team analyzed the available data by person, place, and time and noted that cases ranged in age from 5-85, were approximately 50% male, and the majority (15/22) were from Alberta. 3 cases were hospitalized and there were no deaths. The epi curve showed a peak in cases during the last week of August and first week of September.  

6. Generate hypotheses: What is your best, educated guess about the source of the outbreak?

Descriptive analysis of exposure information, review of the available science, and further interviewing of cases can provide evidence for hypothesis generation.

In our fictional outbreak, OMD conducted case re-interviews using a national hypothesis-generating questionnaire. Data were entered into a database and analyzed. A recent national study was used to establish the expected proportion of Canadians that reported eating food items contained in the questionnaire over the previous seven days, restricted to August and September. Comparisons between case exposure frequencies and population values were made using binomial probability at a 0.05 level of significance. Results of this analysis indicated that any beef and ground beef were reported by cases significantly more frequently than what is usually reported by the general population. Evidence from past outbreaks indicated that ground beef was a plausible source of E. coli O157.    

7. Test hypotheses: Conduct an analytic study to determine if there is a statistically significant association between exposure and outcome.

Although this is not always feasible or necessary, a case-control, cohort, or case-case comparison study maybe useful to build evidence implicating a source for an outbreak.

In our fictional outbreak, the data implicating ground beef were very strong and the investigative team decided that it was not necessary to conduct an analytic study. However, they could have done a case-case comparison with the food exposures of cases in a past E. coli O157 outbreak associated with leafy greens.

8. Implement prevention and control measures: Break the chain of infection.

Points to intervene include at the infectious agent, reservoir, portal of exit, mode of transmission, portal of entry, and susceptible host. In a foodborne illness outbreak, a food safety investigation often occurs in parallel to the epidemiologic investigation. The food safety investigation may include food sampling and traceback on exposures of interest. Sufficient evidence may lead to product action – usually a recall.

In our fictional outbreak, the CFIA conducted sampling of open (from case homes) and closed (from retail) ground beef product. The outbreak strain was found in 1 open and 3 closed samples, and there was sufficient packaging information to trace back the source of the ground beef to one particular plant. A product recall was issued. Prior to the product recall, PHAC issued a Public Health Notice (PHN) warning consumers in Alberta and Saskatchewan of the outbreak link to ground beef.  

9. Communicate findings: Knowledge translation.

Formal or informal mediums for knowledge translation exist and what form is used depends on the audience and the key messages. Recommendations that improve the investigative process should be shared as appropriate. Outbreaks with a novel source, innovative investigative techniques, and a large impact should be published or presented.

In our fictional outbreak, the investigative team created a final investigation summary, which was shared with all investigative partners. They also shared the summary and key investigation indicators on CNPHI to allow other Canadian public health partners to learn from their experiences. Finally, a manuscript is in the works.

10. Debrief team: Discuss what did and did not go well during the course of the investigation.

This is an opportunity to identify processes that work well, as well as those that need improvement. In our fictional outbreak investigation, everything went smoothly and partners decided that a debrief was not needed.

The above example is an idealized version of what may happen in the real world. In reality, many foodborne illness outbreaks are never solved. This is due to factors such as delays in case reporting, poor exposure information, lack of product or packaging details, and absence of leftover product for sampling. Even when a source is identified, the root cause of contamination may remain unknown. However challenging, investigation of foodborne illness outbreaks remains an important function of public health, and successful investigations can prevent excess morbidity and mortality.


Public Health Agency of Canada. (2018). Public health notice – Outbreak of E. coli infections linked to romaine lettuce. Retrieved from html

Public Health Agency of Canada. (2018). Public health notice – Outbreak of Salmonella infections possibly linked to long English cucumbers. Retrieved from services/public-health-notices/2018/outbreak-ecoli-infections-linked-romaine-lettuce.html

Thomas MK, Murray R, Flockhart L, Pintar K, Pollari F, Fazil A, et al. (2013). Estimates of the burden of foodborne illness in Canada for 30 specified pathogens and unspecified agents, circa 2006. Foodborne Pathogens and Disease, 10(7), 639-48.                                                     doi: 10.1089/fpd.2012.1389

Thomas MK, Murray R, Flockhart L, Pintar K, Fazil A, Nesbitt A, et al. (2015). Estimates of Foodborne Illness-Related Hospitalizations and Deaths in Canada for 30 specified Pathogens and Unspecified Agents. Foodborne Pathogens and Disease, 12(10), 820-27.                  doi: 10.1089/fpd.2015.1966

Public Health Agency of Canada. (2011). Pathogen safety data sheets: Infectious substances– Salmonella enterica spp. Retrieved from

Public Health Agency of Canada. (2016). Symptoms of listeriosis (Listeria). Retrieved from

Public Health Agency of Canada. (2017). Surveillance of E. coli (Escherichia coli) infection. Retrieved from /en/public-health/services/diseases/e-coli/surveillance-e-coli.html

Government of Canada. (2018). National Enteric Surveillance Program (NESP) Annual Summary 2016. Retrieved from

Public Health Agency of Canada. (2018). Canada’s Foodborne Illness Outbreak Response Protocol (FIORP): A guide to multi-jurisdictional enteric outbreak response. Retrieved from health-risks-safety/canadas-foodborne-illness-outbreak-response-protocol-fiorp-guide-multi-jurisdictional-enteric-outbreak-response.html

The following two tabs change content below.

Nicole Haywood

Associate editor for the IJHS. Bachelor of Health Sciences, class of 2014, University of Ottawa.

Commentez / Comment: