Ontario documented annual incidence rates of reportable GI cases ranging from 69.2 to 83.9 per 100,000 persons per year for the years 2007 to 2009. Previous studies in Ontario reported on the same diseases, with the exception of the four parasitic diseases (amebiasis, cryptosporidiosis, cyclosporiasis, and giardiasis) included in the current study. As a result, direct comparisons to earlier studies are not easily made. Combined, these four parasitic diseases contributed 2,927, 2,809, and 2,513 cases in 2007, 2008, and 2009, respectively. If these diseases are excluded, the adjusted GI rates would be 61.1, 56.6, and 49.8 per 100,000 persons per year. For the years 1997 to 2003, the annual incidence rate ranged from a high of 89.8 in 1998 to a low of 60.2 per 100,000 persons in 2003 [5, 7]. These rates suggest a slight decrease in the rate of GI in Ontario compared to a decade earlier.
Campylobacteriosis was the most frequently reported GI from 2007 to 2009, making up 36.5% of the total number of cases. The mean annual incidence rate of 28.1 per 100,000 persons for this period was lower than the 42.6 per 100,000 persons reported for the period 1997 to 2001 . Salmonellosis was the second most frequently reported GI from 2007 to 2009, making up 25.1% of the total number of cases. The mean annual incidence rate of 19.4 per 100,000 persons for this period was slightly lower than the 22.6 per 100,000 persons reported for the period 1997 to 2001 . It should be noted that since our study is based on reportable disease data, we could not include viruses that cause GI including the noroviruses, which are estimated to be responsible for 58% of GI caused by known pathogens in the United States, followed by nontyphoidal Salmonella spp. at 11% .
Listeriosis and VTEC-illness had the highest percentage of outbreak-related cases. The wide range in the annual proportion of cases associated with these two diseases was due to significant outbreaks of VTEC-illness in 2007 and 2008, and a large listeriosis outbreak in 2008.
Similar to previous studies, we found that children four years of age and younger have the highest incidence rate for most GI [5–7, 11]. Overall, we found significantly more GI cases in males, also consistent with previous studies [5, 6, 11].
There were pronounced seasonal trends in many of the diseases examined. All of the diseases exhibiting pronounced domestic seasonal patterns in this study increased in the summer; a pattern seen in previous Ontario studies [5–7, 11]. This increase is thought to be due to increased outdoor activities (e.g. recreational water use), social gatherings with food (e.g. barbeques, picnics), an increase in mechanical vectors such as flies, and warmer temperatures that promote pathogen growth [5–7].
Approximately one-quarter of GI cases were acquired outside of the province, a proportion similar to the 24.6% of cases related to travel between 1997 and 2001 . Travel-related seasonal patterns were generally not the same as domestic seasonal patterns. Taylor et al.  also found that international travel had significant impact on the epidemiology of GI in the Canadian province of British Columbia, with the proportion of travel-related cases exceeding that of domestic cases for a group of 13 GI between January and March. This highlights the importance of completing exposure histories so that trends and rates for domestic and travel-related diseases can be assessed separately in order to properly target public health actions .
Source attribution and risk setting
A wide diversity of bacterial, parasitic and viral infections cause GI, each with their specific complex transmission dynamics; GI pathogens can be transmitted from person to person, via contaminated food, water, fomites, as well as through direct or indirect contact with animals. There are a number of different approaches (microbiological, epidemiological, intervention studies and expert elicitation) used to attribute illnesses to specific sources, primarily for “foodborne illnesses”, where sources include animal reservoirs and vehicles (e.g. foods) .
In our study, we used the public health investigator’s determination of most likely setting and source of illness for each case, an approach used in all of the previous Ontario studies [5–7, 11]. This approach to source attribution was also used by Dumoulin et al.  in a region of Ontario, who successfully used standardized questionnaires to determine the most likely source of infection from public health interviews. While the level of evidence for this type of source attribution is considered to be low, we believe these findings are nonetheless useful in describing the exposures identified and investigated by local public health authorities. In our study, investigators reported a known source and setting for 26% and 15% of domestic sporadic cases respectively, with percentages varying by pathogen. The percentage likely reflects the investigators’ confidence in identifying a source. Factors that likely influence this percentage include: the number of days from exposure to investigation and the associated recall bias, the investigators’ bias and knowledge of the sources as well as the transmission modes for the various pathogens, the case’s understanding or bias pertaining to their illness and its cause or source, and the effort made by the investigator. It should be recognized that with these limitations, the source is frequently unknown. Notably, the “known” sources of GI in our study are not systematically supported with information such as positive food samples or a statistical association through using case–control study methods, and are therefore highly susceptible to the various biases described above.
For the approximately one-quarter of cases with a known exposure source, we identified food as the primary exposure source for 54.2% of GI reported in Ontario from 2007 to 2009; food was also the most common exposure source for seven of the diseases examined: botulism, campylobacteriosis, cyclosporiasis, listeriosis, salmonellosis, VTEC-illness, and yersiniosis. In previous Ontario studies the overall proportion of GI attributed to food (when the source was known) was higher, ranging from 73.1% in 2003 to 75.9% in 2002 [5, 6]. However, a more recent study in the Waterloo region of Ontario found the proportion attributed to food to be 57.3% between 1990 and 2004, and 41.2% between 2006 and 2010 . Even if the percentage of foodborne transmission is indeed lower than previously identified, the magnitude is still large enough that continued vigilance of food along the entire “farm-to-fork” continuum is required.
Our finding that contact with animals was an important source of illness accounting for approximately 20% of all GI, was similar to that found by Dumoulin et al. for the years 2006 to 2010 (17.9%) . This proportion was higher than that demonstrated in previous studies, where the proportion ranged from 1.0% in 2003 to 5.8% between 1997 and 2001 [5–7]. Further studies may be warranted to examine the cause of this increasing trend, as well as public health prevention strategies.
For amebiasis, hepatitis A, and shigellosis, person-to-person transmission was the most common primary exposure source. These findings reinforce the need for greater understanding of the importance of personal hygiene practices among cases, their household members, and other close contacts. In addition, more targeted educational information may be warranted for groups that are known to be at higher risk of acquiring these diseases, such as travellers to high-risk areas (e.g. Canadian residents visiting relatives abroad) and men who have sex with men.
The private home was the most common primary exposure setting for nine of the 14 diseases, accounting for less than one-half of all sporadic domestic GI cases. Food premises were identified as the risk setting for approximately one-third of these cases. Similar to the overall incidence rate calculations above, if the four parasitic diseases are removed from the risk setting proportions, private homes and food premises were identified as risk settings for 48.5% and 33.3% of cases, respectively. In previous years, the private home was also the most common overall risk setting for GI cases overall [5–7]. Moreover, when we examined only the cases exposed in the home, we found that food was still the most commonly reported source (62.6%). At 33.3%, the proportion of GI cases thought to have been acquired at a food premise appears to be increasing, as it was 14.1% between 1997 and 2001 , 15.0% in 2002 , and 20.7% in 2003 . The importance of the home and food premises as risk settings for acquiring GI illness also emphasizes the continued need for a farm to fork approach to concurrently reducing pathogens in food as well as reducing risk at the consumer level.
There were many factors that have had an impact on GI case reporting over the last decade, making comparisons over time difficult. Many of these changes may have increased the likelihood of case detection or reporting, improved case management, and ultimately, limited further transmission of disease. A full discussion of these factors is beyond the scope of this paper; however, in Ontario some of the factors included: changes in laboratory testing, changes in case definitions, and a change in the reporting system in 2005 from the Reportable Diseases Information System (RDIS) to iPHIS.
Reportable diseases represent the so-called ‘tip of the iceberg’; only a fraction of GI are reported to public health. While underreporting varies by pathogen, for every GI reported, there are an estimated 10–49 cases in the community that are not reported . Underreporting is a well-known limitation of passive surveillance systems in general; however, since our data are population-based and likely representative, under-reporting will likely only differ slightly over time or for subpopulations, meaning the data can still be used to elucidate epidemiologic trends.
Overall, approximately one-third of cases were not successfully followed up. In public health practice, 100% follow-up is not attainable due to numerous factors, such as cases refusing to be interviewed or public health having incorrect contact information for the patient. Imperfect reporting is a common characteristic of passive surveillance systems such as the one in this study, where the cost of completely capturing data is likely prohibitive and must be balanced against timely capture of enough data to be useful (e.g., to detect outbreaks, vulnerable sub-populations, and trends) . In general, our findings suggest that greater effort was made to follow up pathogens that were considered to have greater morbidity and mortality. Further interpretation of the differences in follow-up between diseases should be done with caution, as there are currently no unified and enforceable procedures for following up reportable diseases across all health jurisdictions in Ontario. Currently, the data collection requirements for jurisdictions are mandated by provincial regulations and infectious disease protocols, and there is no information regarding what a reasonably successful follow-up rate should be by disease.