The results reported here match those from the previous Canadian study , where the incidence rate was 1.3 episodes per person-year, the monthly prevalence was 10%, the average annual probability was 71% and the point prevalence was 1.46%, confirming that the magnitude of gastrointestinal illness at the community level in Canada is considerable and does not appear to vary with geography. These results are also similar to those reported in recent international studies on acute GI in developed countries [7, 9, 12, 17]. However, due to 'telescoping', estimates based on retrospective studies may overestimate results compared to those generated by prospective studies. The study in England found that their retrospective estimate was nearly three times that of their prospective estimate . Likely the point prevalence is a better measure of prevalence as it is less subject to 'telescoping'.
Because the three study regions capture the full spectrum of community types in BC, the results of this study can be reasonably extrapolated to the entire province. In doing so, we estimate roughly 400,000 cases of acute GI per month in BC, with each individual experiencing an estimated 4.8 person-days of acute GI annually. This would translate into 19.2 million days of acute GI experienced in BC each year. Additionally, we would expect over 12 million missed workdays by adults, and over 10 million missed school days among children in BC each year as a result of acute GI. Thus the burden of acute GI in the BC population remains substantial and is associated with discomfort, disability and a significant number of days of lost productivity due to sick time or time taken to care for relatives. In contrast chronic disease surveillance in 2001 identified 574 deaths per 100,000 BC residents per year, though direct comparison with the calculated burden of GI presented here is impossible, this does present a crude value for relative comparison .
Much discussion has occurred around comparing results of studies that have used different case definitions. Currently there is a group of international researchers attempting to resolve this issue [19, 20] This study used a broad case definition for high sensitivity and case capture. Other studies have used more strict definitions for diarrhoea [6, 7, 9, 21–23]. The change resulting from redefining our definition of diarrhoea as three or more loose stools or stools with abnormal liquidity in 24 hours is minimal: the average monthly prevalence would be 8.8% (95% CI 8.0 – 9.6), the average annual probability would be 69.8% (95% CI 66.2 – 73.3) and the incidence rate would be 1.1 (95% CI 1.0 – 1.3) episodes of acute GI per person-year. Even under this stricter definition, the burden remains substantial at over 17 million days of acute GI experienced in BC each year. Further research into the impact of choice of symptomatic case definition is needed, and ultimately a validated, standard definition would be used to ensure comparability between studies.
The age and gender distribution of acute GI in this study is similar to other reported results. Significantly more female than male cases were observed. Females in the 25–64 year old age group had a slightly higher prevalence though not statistically significant; in the previous Canadian study, this difference was significant . Studies in England  and Ireland  also reported an increase in GI among females in the 25–44 year old age group. Higher rates in this adult age group may represent increased exposure of parents with children , and for women in particular, this may also be the result of relatively more exposure in the kitchen during meal preparation (e.g. handling of raw products) .
The temporal distribution of acute GI for individuals in the 10–24 year old age group followed a similar pattern to other studies, with a bimodal distribution peaking in the summer and winter months [6, 13]. This probably reflects the association of bacterial and parasitic gastrointestinal diseases with a summer seasonality [27, 28], and viral gastrointestinal disease with a winter seasonality . However, those in the 0–9 year old age group had a different temporal distribution, reporting peaks in the fall and spring months. This is consistent with rotavirus and norovirus infections, which are common in young children and have been shown to have peaks occurring from, fall to spring [30, 31]. In the previous Canadian study  there was no temporal trend for this age group, and this is likely due to that study's smaller sample size.
A limitation of this study was its low response rate (44.3%); although it is consistent with recent similar international studies [6, 9, 12], and is higher than that in the previous Canadian study (36.6%; ), potentially due in part to the introductory letter . A potential bias associated with low response is non-response bias, which has been discussed in detail elsewhere .
This study was administered by telephone and thus will not capture those who do not own a telephone (1.4% of BC residences do not own at least one telephone or cellular telephone number ), or those who may not be home due to their illness. There is the potential that the rate of illness for those without telephones is different to the rate of illness of those with telephones and thus these results could be biased. For example, if not having a telephone is related to lower income and higher disease rate, then this may underestimate the incidence and prevalence of acute GI in the population, however if not having a telephone is associated with higher income- as may be the case given the emergence of cellular phones- and lower disease rate, this may overestimate the incidence and prevalence of acute GI in the population. As well if individuals were in hospital or institutionalized or had died as a result of their illness, this would not been captured and would thus potentially bias the results, by underestimating the incidence, prevalence and severity of acute GI in the population.