A descriptive analysis of the spatio-temporal distribution of enteric diseases in New Brunswick, Canada

Abstract

Background: Enteric diseases affect thousands of Canadians annually and several large outbreaks have occurred due to infection with enteric pathogens. The objectives of this study were to describe the spatial and temporal distributions of reportable Campylobacter, Escherichia coli, Giardia, Salmonella and Shigella from 1994 to 2002 in New Brunswick, Canada. By examining the spatial and temporal distributions of disease incidence, hypotheses as to potential disease risk factors were formulated.

Methods: Time series plots of monthly disease incidence were examined for seasonal and secular trends. Seasonality of disease incidence was evaluated using the temporal scan statistic and seasonal-trend loess (STL) decomposition methods. Secular trends were evaluated using negative binomial regression modeling. The spatial distribution of disease incidence was examined using maps of empirical Bayes smoothed estimates of disease incidence. Spatial clustering was examined by multiple methods, which included Moran's I and the spatial scan statistic.

Results: The peak incidence of Giardia infections occurred in the spring months. Salmonella incidence exhibited two peaks, one small peak in the spring and a main peak in the summer. Campylobacter and Escherichia coli O157 disease incidence peaked in the summer months. Moran's I indicated that there was significant positive spatial autocorrelation for the incidence of Campylobacter, Giardia and Salmonella. The spatial scan statistic identified clusters of high disease incidence in the northern areas of the province for Campylobacter, Giardia and Salmonella infections. The incidence of Escherichia coli infections clustered in the south-east and north-east areas of the province, based on the spatial scan statistic results. Shigella infections had the lowest incidence rate and no discernable spatial or temporal patterns were observed.

Conclusions: By using several different spatial and temporal methods a robust picture of the spatial and temporal distributions of enteric disease in New Brunswick was produced. Disease incidence for several reportable enteric pathogens displayed significant geographic clustering indicating that a spatially distributed risk factor may be contributing to disease incidence. Temporal analysis indicated peaks in disease incidence, including previously un-reported peaks.

Fig. 1

Monthly incidence of reportable enteric illness (1994–2002) in New Brunswick, Canada. Rates are expressed in cases per 100,000 person-month. Data are presented as 2 month moving averages (red-dashed) and 13 month (blue-dashed) moving averages

Fig. 2

Seasonal decomposition of the monthly incidence reportable enteric illness (1994–2002) in New Brunswick, Canada

Fig. 3

Empirical Bayes smoothed incidence rates for reportable enteric illness (1994–2002) in New Brunswick, Canada at the Census Subdivision level (CSD). Darker areas indicate areas of higher disease incidence

Fig. 4

Empirical Bayes smoothed incidence rates for reportable enteric illness (1994–2002) in New Brunswick, Canada at the watershed level (WS). Darker areas indicate areas of higher disease incidence

Fig. 5

Significant spatial clusters (p-value < 0.05) for reportable enteric illness (1994–2002) in New Brunswick, Canada at the Census Subdivision level (CSD) as identified by the spatial scan statistic

Fig. 6

Significant spatial clusters (p-value < 0.05) for reportable enteric illness (1994–2002) in New Brunswick, Canada at the watershed level (WS) as identified by the spatial scan statistic